{ "schema_version": 1, "reviewed_at": "2026-10-07", "coverage_complete": false, "note": "Deduplicated retrieval candidates, not full-text-reviewed or accepted implementations. Track status is the pre-batch repository lookup; latest decisions are in the discovery ledger.", "candidates": [ { "arxiv_id": "2610.00015", "title": "From Proposal to Verified Effect: Praxa, an Evidence-Bound Harness for Governed AI Agent Execution", "published": "2026-07-27T05:57:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00025", "title": "Measuring the Microtask Eligibility Gap: When Is an Off-the-Shelf SLM Enough for an Agent Harness?", "published": "2026-09-01T21:07:51Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00035", "title": "Integrating Fairness and Explainability in a Multiple Instance Reinforcement Learning System", "published": "2026-09-02T17:30:18Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.00047", "title": "Characterizing a Configuration Where Inference-Time PRM-Pruned Fragment Grafting Is Inert: Evidence from Three Reasoning LMs", "published": "2026-09-03T11:06:07Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00054", "title": "The First Token Is Not the Verdict: Hidden Costs of Reading LLM Judges Without Generating", "published": "2026-09-04T00:45:28Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.00069", "title": "A Framework for Egocentric and Exocentric Procedural Understanding via Temporal Segmentation and Semantic Abstraction", "published": "2026-09-04T17:34:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00084", "title": "Scientific Agents: Evaluating Profession-Specific System Prompts on Scientific Tasks", "published": "2026-09-07T00:21:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00093", "title": "Safety in Self-Evolving Agents: A Survey", "published": "2026-09-08T05:04:40Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00096", "title": "FACET at WMT 2026 Automated Translation Quality Evaluation Task", "published": "2026-09-08T14:49:27Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00097", "title": "DramaAgent: Agentic Storytelling Video Generation", "published": "2026-09-08T16:11:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00106", "title": "SyntheticHLS: Building Diverse Synthetic High-Level Synthesis Datasets using LLMs", "published": "2026-09-09T02:40:29Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00182", "title": "Localizing Post-Wire Semantic Changes in MCP Agent Frameworks", "published": "2026-09-17T04:06:38Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00188", "title": "Uncertainty-Aware RL-Controlled Adaptive 3D Mapping", "published": "2026-09-17T17:53:51Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00196", "title": "GPEC: Efficient Pre-LLM Gaussian Process Embedding Correction for Cardiac Video Caption Generation", "published": "2026-09-18T19:52:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00197", "title": "Comedic Fool's Gold: Reward Exploits and Countermeasures in Conversational Humor", "published": "2026-09-18T20:14:09Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00223", "title": "A Holistic Assessment of the Carbon Footprint of Noor, a Very Large Arabic Language Model", "published": "2026-09-22T11:25:39Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00251", "title": "The Null Is the Hard Part: Exact Tests for Memorization in Generative Models", "published": "2026-09-23T17:59:25Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.00253", "title": "System Attribution in LLM Brand Recommendations: Single Responses Identify the System, Aggregated Brand Profiles Do Not Transfer", "published": "2026-09-23T21:57:07Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.00256", "title": "Verification Pulses and the Cost of Escaping Wrong Consensus", "published": "2026-09-24T03:29:27Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00313", "title": "Rules to Tools: Executable Checks for LLM Agents in Scientific Computing", "published": "2026-09-29T00:36:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00316", "title": "DuplexSpeechBench-Document Grounding: Benchmarking Document Grounding and Hallucinations in Voice Agents", "published": "2026-09-29T01:43:55Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00317", "title": "DriftOPD: Sequence-Level Reverse-KL Distillation for One-Step VLA Policies", "published": "2026-09-29T02:55:12Z", "status_at_scan": "implemented", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00332", "title": "The Weakest Link: Distilling LLM Reasoning with Worst-Case Constrained Reinforcement Learning", "published": "2026-09-29T13:51:52Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00333", "title": "LEGO-OPD: Factorized Teacher Composition for Multimodal On-Policy Distillation", "published": "2026-09-29T13:57:51Z", "status_at_scan": "implemented", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00354", "title": "Proof-Gated Signing: Solver-Checked Transaction Guards that Hold Under State Drift for Onchain AI Agents", "published": "2026-09-30T00:24:01Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00363", "title": "Deep Learning for Anomaly Detection in Railway Systems: A Structured Survey", "published": "2026-09-30T06:29:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00367", "title": "MoRA: MoE Pruning via Router Bias Learning and Expert Approximation", "published": "2026-09-30T06:38:43Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00372", "title": "When Harnesses Lose the Signal: Causal Evaluation of Recovery in LLM Agents", "published": "2026-09-30T07:04:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00374", "title": "Faithful Chart Generation for Multimodal Deep Research: Frame-Evidence Co-Adaptation", "published": "2026-09-30T07:33:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00385", "title": "FAER: Auditable Utility-Aligned Trajectory Replay for Language Model Post-Training", "published": "2026-09-30T09:41:08Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00388", "title": "T2SPO: Trajectory-to-Step Policy Optimization for Agentic Reinforcement Learning", "published": "2026-09-30T10:34:27Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.00398", "title": "WIPSNet: Deep Learning for Paediatric Wheeze Detection from Overnight Impedance Pneumography", "published": "2026-09-30T11:58:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00399", "title": "Metacognitive Reasoning in Energy Based Models using Instance Based Learning Theory", "published": "2026-09-30T12:08:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00400", "title": "Representation Transitions Reveal Emerging Safety Risks in Multi-Turn LLM Agents", "published": "2026-09-30T12:30:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00406", "title": "LLM-as-a-Judge for Low-Resource Languages: Adapting Ragas and Comparative Ranking for Romanian", "published": "2026-09-30T13:19:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00425", "title": "Code That Works, Environments That Don't: Measuring Environment Reproducibility in AI-Generated Software", "published": "2026-09-30T16:04:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00426", "title": "IrekoGPT: Turning Structured Pruning into Post-Hoc Slimmable LLMs", "published": "2026-09-30T16:08:59Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00430", "title": "Memetic Trojans: Social Contagions as Carriers of Adversarial Payloads in Agent Networks", "published": "2026-09-30T16:43:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00431", "title": "ChainLoRA: Geometry-Preserving Task Vector Merging for Continual Learning in LLMs", "published": "2026-09-30T16:43:34Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00432", "title": "XOR-Trellis: Ultra-Low-Complexity Dequantization and Curvature-Aware Hadamard-Free LLM Quantization", "published": "2026-09-30T17:05:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00435", "title": "How AI Agents Discover Scientific Equations: From Hydrotope Rediscovery to New Water-Wave Amplitudes", "published": "2026-09-30T17:16:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00437", "title": "JevSpawn: Adaptive Agentic Inference through Compositional Action Spaces", "published": "2026-09-30T17:23:01Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00465", "title": "AIR-LLM: Broadcasting AI Weights over Radio for Memory-Free Edge LLM Inference via RF Computing", "published": "2026-09-30T18:00:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00487", "title": "ScaffoldM3C: A Multimodal Sequential Monte Carlo Framework for Generative Stable Construction Planning", "published": "2026-09-30T18:00:17Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00497", "title": "Gumbel Straight Flow: Distilling Autoregressive Models into One-step Flow Maps", "published": "2026-09-30T18:01:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00499", "title": "Denoising Surface: Modeling and Predicting Inference Cost for Diffusion LLM Serving", "published": "2026-09-30T18:01:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00511", "title": "Before Agents Decide: Epistemic Action in LLM-Based Systems", "published": "2026-09-30T18:07:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00540", "title": "Assessing the Impact of Language Disparity on Multilingual Linguistic Ability in Large Language Models", "published": "2026-09-30T18:22:47Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00541", "title": "Random Recursive Models", "published": "2026-09-30T18:23:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00557", "title": "No One Architecture Fits All: A Cross-Environment Evaluation of Hierarchical Red Team Agents", "published": "2026-09-30T18:32:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00559", "title": "PhysVista: Benchmarking Physical Intelligence in VLMs via a Perception-Reasoning-Assessment Loop", "published": "2026-09-30T18:34:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00562", "title": "Can LLMs Reason Over Long Horizons? An Empirical Evaluation of Context Strategies for Longitudinal Clinical Reasoning", "published": "2026-09-30T18:36:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00568", "title": "Emergent Unfaithfulness: How Alignment Training Causes Language Models to Silently Override Task Faithfulness", "published": "2026-09-30T18:41:05Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00573", "title": "FORTE: Adaptive Scoring and Exact Keyframe Selection for Long-Video Question Answering", "published": "2026-09-30T18:45:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00574", "title": "Make Sparse Rewards Count: Density-Aware Reward Aggregation for Multi-Reward RL", "published": "2026-09-30T18:45:26Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00590", "title": "Towards Hierarchical Cyber Defense with Large Language Models: From Planning to Execution", "published": "2026-09-30T18:53:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00592", "title": "ALER: Adaptive Learnable Experience Rewriting for Reinforcement Learning", "published": "2026-09-30T18:55:45Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.00604", "title": "MIKASA-Robo-VLA: Benchmarking Memory in VLA Models for Long-Horizon Manipulation", "published": "2026-09-30T19:10:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00609", "title": "Legal Research Bench: Measuring End-to-End Reliability in Long-Horizon Legal Research Agents", "published": "2026-09-30T19:13:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00610", "title": "Explainable Suicide Risk Assessment on Social Media with Multi-Task QLoRA", "published": "2026-09-30T19:13:19Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00613", "title": "Spatial Strategies, Not Actions: Vector-Quantized Geodesics as Tools for LLM-Driven Agents", "published": "2026-09-30T19:16:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00619", "title": "Beyond Supra-Competitive Outcomes: Collusive Behaviour in Deep Reinforcement Learning for Optimal Execution Games", "published": "2026-09-30T19:20:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00622", "title": "Understanding and Mitigating Library-Related Issues in LLM-Generated Code", "published": "2026-09-30T19:25:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00623", "title": "HAWK: Rethinking Multimodal Drafting for Speculative Decoding", "published": "2026-09-30T19:26:06Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00629", "title": "ASAD: Adaptive Software Agents for Debugging", "published": "2026-09-30T19:32:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00636", "title": "CompMat-Bench: Benchmarking AI Agents for Computational Materials Science", "published": "2026-09-30T19:37:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00648", "title": "Incident-Arena: Getting agents to the last nine of reliability", "published": "2026-09-30T19:50:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00650", "title": "Self-Evolving Coding Rules for AI Coding Agents", "published": "2026-09-30T19:51:41Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00651", "title": "Agent Evaluation Reliability: More Tasks Won't (Always) Fix An Agent Leaderboard", "published": "2026-09-30T19:52:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00654", "title": "When More Data Is Not Enough: The Context-Sufficiency Frontier in Generative AI Personalization", "published": "2026-09-30T19:54:26Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.00661", "title": "Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models", "published": "2026-09-30T19:59:48Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00663", "title": "Backdoor Containment via Expert Quarantine and Shutdown in LLMs", "published": "2026-09-30T20:02:14Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00665", "title": "Analysis of Quantized and Efficiently Adapted Protein Language Models", "published": "2026-09-30T20:02:41Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00666", "title": "VisionQ: VLM-as-a-Judge Taxonomy, Dataset and Benchmark for Qualitative Analysis in Computer Vision", "published": "2026-09-30T20:03:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00673", "title": "Closing the Loop: Practical Training Recipes for Looped Language Models", "published": "2026-09-30T20:09:58Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00675", "title": "LabBook: Harnessing Experimental History for Efficient LLM-Driven Discovery", "published": "2026-09-30T20:11:51Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00677", "title": "Harnessing Vision-Language Models for Perceptual Quality Assessment and Autonomous Content Adjustment in Augmented Reality", "published": "2026-09-30T20:13:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00680", "title": "Curvature Under Attack in hZACH-ViT: Gauge Symmetry, Boundary Saturation, and Adversarial Failure", "published": "2026-09-30T20:20:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00694", "title": "How Divergence Becomes Decision Flips in Compressed Language Models", "published": "2026-09-30T20:37:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00705", "title": "Meta-Multi-Agent Reinforcement Learning for Fast Adaptation of Interactive Policies with Applications to Autonomous Driving", "published": "2026-09-30T20:52:19Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00710", "title": "ReLiveGym: Evaluating Long-Lived Agents over Weeks of Replayed Reality", "published": "2026-09-30T20:59:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00717", "title": "Sequential Functional Structured Tucker Compression for Large Language Model Attentions", "published": "2026-09-30T21:06:29Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00758", "title": "Scalable Multi-Task Inverse Reinforcement Learning", "published": "2026-09-30T21:52:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00767", "title": "Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints", "published": "2026-09-30T22:02:46Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00779", "title": "Effective Synthetic Data Curation Requires Group-Level Signals", "published": "2026-09-30T22:16:23Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00795", "title": "Can large language models unlock discrete data in ophthalmic diagnostic reports?", "published": "2026-09-30T22:35:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00809", "title": "Paying for Too Many Tokens? Valid and Cost-Efficient Multimodal LLM Annotation with Simple Heuristics", "published": "2026-09-30T22:59:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00825", "title": "Align Then Reason: A Multimodal Lip-Sync Judge for Dubbing", "published": "2026-09-30T23:35:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00838", "title": "SHARPO: Segment-Level Credit Assignment for Agentic Reinforcement Learning", "published": "2026-09-30T23:51:11Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.00848", "title": "Geometric Similarity in VLM Low-Level Vision Representations", "published": "2026-10-01T00:06:17Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00849", "title": "Learning Multiple Timescales for Goal-Conditioned Reinforcement Learning", "published": "2026-10-01T00:06:57Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00852", "title": "Child-Adapted Structured Phonological Representations for Interpretable Speech Sound Analysis", "published": "2026-10-01T00:12:27Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00855", "title": "Lang3DSeg: Annotation-Free Open-Vocabulary 3D Segmentation with Point Transformers", "published": "2026-10-01T00:15:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00864", "title": "Kinematic MeanFlow: One-Step Action Generation Policy for Robotic Foundation Models", "published": "2026-10-01T00:31:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00870", "title": "An Educator-Guided LLM Pedagogical Agent for Scaffolded Feedback in Conceptual Database Design", "published": "2026-10-01T00:38:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00872", "title": "MemFit: Efficient Long-Term Agentic Memory", "published": "2026-10-01T00:44:33Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00878", "title": "UniTrackPLA: Unified Panorama-Language-Action Model for Instruction-Guided Navigation and Dynamic Person Tracking", "published": "2026-10-01T00:53:07Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00881", "title": "Machine Translation for Sign Languages", "published": "2026-10-01T00:56:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00883", "title": "DeBERTa-ConPara: Attack-Aware and Deployment-Realistic Detection of AI-Generated Text", "published": "2026-10-01T00:57:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00885", "title": "FORALL-LEAN-AGENT for Auditable Reasoning in Formal Mathematics and Software Verification", "published": "2026-10-01T01:00:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00888", "title": "Match the Distribution, Not the Compute: Post-Training Multi-Token Prediction Heads", "published": "2026-10-01T01:09:18Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00890", "title": "Cross-Benchmark Transfer from RL on Agentic Coding Tasks", "published": "2026-10-01T01:12:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00898", "title": "When Do Biological Reasoning Models Use Their Biological Inputs?", "published": "2026-10-01T01:27:31Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00899", "title": "TOAST: Stochastic Robot Action Tokenization for Autoregressive Vision-Language-Action Models", "published": "2026-10-01T01:28:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00903", "title": "Sharpen Before You Adapt: Data-Free Entry-State Sharpening for Test-Time Reinforcement Learning", "published": "2026-10-01T01:33:30Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00905", "title": "Understanding Issues, Causes and Solutions in Open-Source LLM-based Multi-Agent Systems", "published": "2026-10-01T01:34:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00906", "title": "ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization", "published": "2026-10-01T01:36:03Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00917", "title": "Finding the Right Fit: Model-Harness Interactions across Agent Tasks", "published": "2026-10-01T01:49:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00922", "title": "EyeTAG: Eye Trajectory-Aware Gaze Estimation", "published": "2026-10-01T01:57:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00952", "title": "A Matched-Budget Audit Framework for Recaptioned Image-Text Supervision Distributions", "published": "2026-10-01T02:35:20Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00954", "title": "Beyond Leaderboards: Tokenomics of Agentic Small Language Model Ensembles", "published": "2026-10-01T02:37:56Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.00958", "title": "Role-aware Heuristic Episodic Attention for Conversational LLMs", "published": "2026-10-01T02:39:52Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00964", "title": "RPTune: Learned Context Curation for LLM Catalog Search", "published": "2026-10-01T02:52:24Z", "status_at_scan": "implemented", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00969", "title": "A Citation-Grounded Benchmark for Trustworthy Earnings Call Transcript Analysis with Large Language Models", "published": "2026-10-01T02:59:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00972", "title": "VeriHarness: Scaling Agentic Verification for Long-Horizon Tasks", "published": "2026-10-01T03:02:17Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00973", "title": "Concept Driven Domain Adaptation: Finding an Abstract Needle in a Haystack", "published": "2026-10-01T03:03:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00977", "title": "ABSENTIA: Detecting Broken Access Control Vulnerabilities in Web Applications", "published": "2026-10-01T03:11:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00979", "title": "RISED: RubrIcs for agentic multi-environment Selection and sElf-Distillation", "published": "2026-10-01T03:16:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.00981", "title": "NarrativeFlow: Flow-Based Vision-Language-Action Model Using Robot Velocity Fields", "published": "2026-10-01T03:18:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.00983", "title": "The Devil Is in the Reconstruction Loss Scale: Rethinking Optimization in LLM Quantization", "published": "2026-10-01T03:19:52Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.00984", "title": "HADRec: A Hierarchy-Aware Drug Recommendation Framework by Fusing Molecular Knowledge and Electronic Health Record", "published": "2026-10-01T03:20:54Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.00988", "title": "Auditable Algebraic Counting Field for Cryptic-Pocket Detection from Apo Structures", "published": "2026-10-01T03:22:43Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.00991", "title": "Adapter Thickets: Splitting an RLVR Budget Beats Concentrating It", "published": "2026-10-01T03:25:40Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01017", "title": "Pay for the Fault, Not the Flow: Label-Free In-Flow Multi-Agent Workflow Optimization", "published": "2026-10-01T04:01:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01023", "title": "Groundability, Not Scale Alone: When Weak Reviewers Can Audit Strong Coding Agents", "published": "2026-10-01T04:10:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01026", "title": "It Takes Workflows to Evolve Better Workflows", "published": "2026-10-01T04:15:49Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.01027", "title": "LawCompass: Navigating from Legal QA to Multi-Agent Deep Research with Grounded Evidence", "published": "2026-10-01T04:16:08Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01037", "title": "SLIM: Simplex-Lattice Interpolation Merging", "published": "2026-10-01T04:23:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01042", "title": "Beyond Final Accuracy: Auditing Communication in LLM Multi-Agent Systems", "published": "2026-10-01T04:34:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01054", "title": "Capturing In-Context Learning Dynamics with Task Operators", "published": "2026-10-01T04:50:17Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01066", "title": "Probe with Participation Trophies: Random-Reward RL as a Probe of LLM Capability", "published": "2026-10-01T05:05:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01073", "title": "Safety Must Survive Self-Improvement: Why Failures Persist and How Agents Recover", "published": "2026-10-01T05:13:51Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01076", "title": "GLoC-EHR: Evidence-Cited Clinical Reasoning over Global Context and Local EHR Events", "published": "2026-10-01T05:16:20Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01093", "title": "OrbitTAMP: Grounding Language Models for Task and Motion Planning in Spacecraft Rendezvous", "published": "2026-10-01T05:37:38Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01102", "title": "MASkillBlender: Decentralized Whole-Body Coordination for Multi-Humanoid Loco-Manipulation via Skill Blending", "published": "2026-10-01T05:42:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01127", "title": "Counting and Min-Cost Encoding for Tokenization in Large Language Models", "published": "2026-10-01T06:11:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01128", "title": "Grounding Large Language Models in DSGE Simulators for Policy Generation and Forecasting", "published": "2026-10-01T06:11:48Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01138", "title": "Auditing Action Settlement in LLM Agent Environments: Order, Progress, and Replay", "published": "2026-10-01T06:17:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01161", "title": "My FAULT: Self-Diagnosis as Credit Assignment in Self-Evolving Agentic Reinforcement Learning", "published": "2026-10-01T06:41:10Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.01166", "title": "CineMR: Tool-Integrated Vision-Language Reasoning for Quantitative Cardiac MRI Assessment", "published": "2026-10-01T06:43:42Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01172", "title": "Learning Rate Transfer for Hybrid Transformer-SSM Architectures", "published": "2026-10-01T06:46:03Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01180", "title": "Skeleton-and-Strategy Prompting: Training-Free Negation Understanding for Vision-Language Models", "published": "2026-10-01T06:53:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01184", "title": "ReCast: Contract-Preserving Protection for Fixed-Interface Multimodal Reasoning", "published": "2026-10-01T07:00:48Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01192", "title": "FlashBack: Knowing When to Remember in Streaming Vision-Language Models", "published": "2026-10-01T07:05:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01195", "title": "Federated Agent Optimization", "published": "2026-10-01T07:06:00Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01207", "title": "Dependency-Aware Reward Shaping for Agentic Reinforcement Learning", "published": "2026-10-01T07:16:28Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01234", "title": "ASCRIBE: Atomic and Significance-Based Reasoning for Thai Clinical SOAP Note Generation", "published": "2026-10-01T07:35:30Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01238", "title": "Mixture-Trained Merging for Unified Multi-Objective Models", "published": "2026-10-01T07:37:37Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01243", "title": "When the Judge Acts: Auditing VLM-Guided Image Selection on Culturally Situated Prompts", "published": "2026-10-01T07:40:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01256", "title": "DeFA: Dependency-Guided Failure Attribution for LLM Agents", "published": "2026-10-01T07:50:22Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01257", "title": "Science Utopia? Closed-Loop LLM Simulation of Academic Research Ecosystems", "published": "2026-10-01T07:51:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01270", "title": "Not All Is Lost: Repairing Lossy User Preference States of Personalization Encoders", "published": "2026-10-01T08:09:37Z", "status_at_scan": "implemented", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01275", "title": "Know When to Hold 'em: Correct-Token Retention in Uniform-State Diffusion Language Models", "published": "2026-10-01T08:13:50Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01278", "title": "SCOPE-AD: Sequential cost-aware ordinal-belief planning with energy-based models for diagnostic agents", "published": "2026-10-01T08:16:07Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.01286", "title": "Dyna3: VLM-Guided Training-Free 4D Reconstruction via Depth Foundation Models", "published": "2026-10-01T08:24:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01316", "title": "What Wins a Vote? Formatting, Length, and Lexical Diversity in the French Compar:IA LLM Arena", "published": "2026-10-01T08:47:35Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01325", "title": "PPO-HRAP: Proximal Policy Optimization with a Hybrid Regime-Aware Policy for Risk-Controlled Trading", "published": "2026-10-01T08:49:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01345", "title": "ARCCS: An Automated Regulatory Compliance Checking System", "published": "2026-10-01T09:17:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01349", "title": "PACE: Provenance-Aware Capability Enforcement for Tool-Using LLM Agents", "published": "2026-10-01T09:18:05Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01364", "title": "LLM-Driven Multi-Agent Control for Skill-Based Smart Manufacturing", "published": "2026-10-01T09:34:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01389", "title": "AiSearch: Interactive Multi-Modal Search with VLMs", "published": "2026-10-01T09:56:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01395", "title": "AF-Muon: An AdamW-Free Muon Optimizer for Tied-Embedding Models", "published": "2026-10-01T09:59:36Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01415", "title": "Beyond Memory: Harnessing Long-Horizon Agents with Explicit Belief States", "published": "2026-10-01T10:21:46Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01434", "title": "MWOP: Modality-aware Width-wise Operation Pruning for Efficient MLLMs", "published": "2026-10-01T10:31:21Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01451", "title": "A Multi-Agent LLM Framework for Personalized Health Checkup Interpretation and Guidance", "published": "2026-10-01T10:48:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01490", "title": "The Persona Is Still There, but Who Is Speaking? Latent Identity Reversion in Persistent AI Agents", "published": "2026-10-01T11:30:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01491", "title": "Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories", "published": "2026-10-01T11:31:37Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01492", "title": "Q-SPT: Learnable Query-Based Compression for Low-Frame-Rate Speech Tokenization", "published": "2026-10-01T11:35:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01496", "title": "SALD: Self-Referenced Advantage Learning for Diffusion Models", "published": "2026-10-01T11:37:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01499", "title": "VTR-Bench: A Systematic Benchmark for Evaluating Visual Text Rendering in Video Generation", "published": "2026-10-01T11:40:03Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01506", "title": "MCRI: A Four-Dimensional Framework for Analyzing and Evaluating Agent Skills", "published": "2026-10-01T11:43:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01508", "title": "OverAct: Measuring and Mitigating Proactive Over-Authorization in LLM Tool-Calling Agents", "published": "2026-10-01T11:44:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01509", "title": "Sharpening Tax in Post-Training", "published": "2026-10-01T11:46:00Z", "status_at_scan": "implemented", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.01511", "title": "GAW-PO: Preference Optimization with Gradient-Aligned Token Weights", "published": "2026-10-01T11:46:28Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01513", "title": "Decision Titan: Test-Time Training for Long-Term Memory in Offline Reinforcement Learning", "published": "2026-10-01T11:48:32Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.01514", "title": "How the Audit Rule Shapes Faithful Factor Explanations in LLMs", "published": "2026-10-01T11:48:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01530", "title": "Calibrating Prediction Timeliness Through Multi-Objective Hyperparameter Optimization for Remaining Useful Life Prediction", "published": "2026-10-01T12:06:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01531", "title": "Towards Reliable Vision-Language Models for Autonomous Driving", "published": "2026-10-01T12:06:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01533", "title": "Neither Black nor White: Balancing Semantic and Collaborative Signals with Graph-Informed Semantic IDs (GrIS)", "published": "2026-10-01T12:08:56Z", "status_at_scan": "implemented", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01537", "title": "FedFit: Federated Fine-Tuning of LLMs via Vector-Bank Parameterization and Quantization", "published": "2026-10-01T12:09:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01548", "title": "Range-GRPO: Policy Optimization via Pairwise Relations among Reward Intervals", "published": "2026-10-01T12:16:34Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01560", "title": "AURAL: Adaptive Latent Reasoning with Joint Chunk for Speech Language Models", "published": "2026-10-01T12:28:03Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01564", "title": "Chaining Skills to Hijack LLM Agents", "published": "2026-10-01T12:28:49Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01569", "title": "Managing Context and Communication in Distributed Agentic UAV Swarms", "published": "2026-10-01T12:30:29Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.01578", "title": "The hidden advantage of mask resampling: a theory of masked autoencoders", "published": "2026-10-01T12:35:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01595", "title": "Before It Fades: Reinforcing Temporal Representations at Inference Time in VideoLLMs", "published": "2026-10-01T12:45:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01605", "title": "Hob-VL: A Benchmark for Visually Grounded Boolean Reasoning", "published": "2026-10-01T12:48:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01625", "title": "Beyond Domain-Level Adaptation: Margin-Oriented Semantic-Appearance Interaction Correction for Personalized Federated Vision-Language Models", "published": "2026-10-01T12:57:30Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01637", "title": "Fusing Visual and Textual Representations via Multi-layer Fusing Transformers for Vietnamese Visual Question Answering", "published": "2026-10-01T13:04:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01640", "title": "Not All Error Yields to Scale: Where Scaling Stops in Vision-Language Inference", "published": "2026-10-01T13:05:39Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01649", "title": "CrossGMN: Graph Metanetworks for Cross-Architecture Weight-Space Transformations", "published": "2026-10-01T13:13:55Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01670", "title": "Do MLLM Judges Judge the Edit? Auditing Bias in Image Editing Evaluation with Verified Quality Preservation", "published": "2026-10-01T13:30:07Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.01687", "title": "Architectural Sampling: Test-Time Scaling via Computational Diversity in Frozen Vision-Language Models", "published": "2026-10-01T13:44:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01705", "title": "AgentWebRec: Compact Evidence Fusion over the Agent Web for Personalized Recommendation", "published": "2026-10-01T13:51:13Z", "status_at_scan": "implemented", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01710", "title": "CoEvolve: Construct-to-Edit Visual Grounding with Bidirectional State Refinement", "published": "2026-10-01T13:54:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01741", "title": "ATI-VLA: Action-Centric Predictive Vision-Language-Action Models via Actionable Alignment Then Adaptive Injection", "published": "2026-10-01T14:12:23Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01746", "title": "End-to-End Learning vs. Modular Architectures: Comparative Insights into Autonomous Driving Systems", "published": "2026-10-01T14:15:06Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01754", "title": "Cog-VADU: A Training-Free Cognitive Reasoning Framework for Video Anomaly Detection and Understanding", "published": "2026-10-01T14:18:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01766", "title": "VideoEvolve: Evolving Agent Harnesses for Video Temporal Grounding", "published": "2026-10-01T14:24:24Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01769", "title": "CONTRA: Discovering and Qualifying Behavior-Changing Questions for Selective Clarification in LLM Code Generation", "published": "2026-10-01T14:26:21Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01785", "title": "VETO: Video Efficient Token Optimization for Vision Language Models", "published": "2026-10-01T14:32:43Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01794", "title": "Continuous Conditioning of VLAs with Augmenting EMG and Visual Task Descriptors", "published": "2026-10-01T14:37:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01815", "title": "Debias Anything: Fairness with Diversity without Supervision in Diffusion Models", "published": "2026-10-01T14:51:03Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.01821", "title": "Beyond Linear Concepts: Discovering and Aligning Non-Linear Concept Manifolds in Large Language Models", "published": "2026-10-01T14:55:03Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01828", "title": "The Asymptotics of Language Model Alignment with Memory", "published": "2026-10-01T15:04:05Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01833", "title": "Continuous Process-Level Evaluation for Evolving Enterprise AI Agent Skills", "published": "2026-10-01T15:06:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01835", "title": "Varda-single-1.0: deterministic data-driven weather forecasting at 1 km resolution over Switzerland's complex topography", "published": "2026-10-01T15:07:20Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01846", "title": "Beyond Decodability: Do Acoustic Factors Drive Predictions in Speech-Based Alzheimer's Assessment?", "published": "2026-10-01T15:13:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01882", "title": "Flowing Faster to Coordinate: One-Step Online Multi-Agent Flow Policies", "published": "2026-10-01T15:36:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.01892", "title": "Selection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents", "published": "2026-10-01T15:42:14Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.01894", "title": "A foundation for systematic analysis of transformers and RNNs for tractography", "published": "2026-10-01T15:43:32Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.01896", "title": "Asynchronous LLM Post-Training: Group-Mass Capping and Convergence Analysis", "published": "2026-10-01T15:45:08Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01908", "title": "Same Reward, Different Skills: When Multimodal RL Learns to Look", "published": "2026-10-01T15:53:22Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.01917", "title": "MoLE: Mixture of Latent Experts for Complementary Visual Reasoning", "published": "2026-10-01T15:54:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01921", "title": "Cross-Lingual Alignment for Decoder-Only Models using MoE Routers", "published": "2026-10-01T15:56:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01939", "title": "Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens", "published": "2026-10-01T16:07:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01944", "title": "Anti-Persona: Disrupting Unauthorized Identity Binding and Recognition in Personalized Vision--Language Models", "published": "2026-10-01T16:10:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01955", "title": "Do Your Own Research: Learning to Forecast by Learning to Search", "published": "2026-10-01T16:12:58Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.01962", "title": "SIEVE: Selective attention-value Suppression for Vision-Language Models Unlearning", "published": "2026-10-01T16:16:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01973", "title": "Token-Level Video Reinforcement Learning", "published": "2026-10-01T16:22:41Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01981", "title": "Universal interpolation for deep residual self-attention networks", "published": "2026-10-01T16:25:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.01999", "title": "From Reasoning Failures to Composable Video Spatial Intelligence", "published": "2026-10-01T16:31:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02002", "title": "Mem++: Non-Destructive Memory for Long-Term Organizational LLM Agents", "published": "2026-10-01T16:32:38Z", "status_at_scan": "implemented", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02005", "title": "Counting Moves, Weighing Voices: Bayesian Dialectical Argumentation for Calibrated Multi-LLM Councils under Persistent Adversaries", "published": "2026-10-01T16:34:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02012", "title": "Bellman Meets Lyapunov: Unsupervised Reinforcement Learning via Mastering Chaos", "published": "2026-10-01T16:38:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02015", "title": "On Language Drift during RLVR Post-Training", "published": "2026-10-01T16:39:48Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02019", "title": "Controllable Multi-label Video Safety Detection via Adaptive Tversky Policy Optimization", "published": "2026-10-01T16:42:30Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.02021", "title": "Task-Adaptive Grounded 3D-Programmers Using 2D VLMs", "published": "2026-10-01T16:43:17Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02038", "title": "Mimir: Physics-Grounded LLM Agents for Long-Horizon Irrigation Control", "published": "2026-10-01T16:50:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02039", "title": "CARM: Cancellation-Aware Response Masking for LLM Reinforcement Learning", "published": "2026-10-01T16:51:56Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02045", "title": "Form and Void: Entangled Composition through an Autonomous AI Agent", "published": "2026-10-01T16:55:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02057", "title": "Optimizing Effective Training Time for Large-Scale Recommendation Systems", "published": "2026-10-01T17:02:16Z", "status_at_scan": "implemented", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.02074", "title": "Homomorphic Advantage Operator: Stabilizing Reinforcement Learning Under Fully Homomorphic Encryption Constraints", "published": "2026-10-01T17:14:32Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02076", "title": "LLM-as-Jev: LLMs Are Already Jev-Style Decision Models -- When and How to Fine-Tune Them", "published": "2026-10-01T17:15:24Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02091", "title": "GeoLatent: Geometry-Guided Latent Structuring with Routed Optimization for 3D Reasoning", "published": "2026-10-01T17:22:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02117", "title": "Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes", "published": "2026-10-01T17:34:10Z", "status_at_scan": "implemented", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.02140", "title": "Finetuning with Sampling: SFT Learns Better Than You Think", "published": "2026-10-01T17:45:07Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02142", "title": "Keyword Harnesses Fail Open: A Cheap Diagnostic Ladder for Tool-Use Claims in Small Language Models", "published": "2026-10-01T17:46:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02144", "title": "Faynt: Scaling and Optimizing Policies for Competitive Melee", "published": "2026-10-01T17:48:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02148", "title": "Omni-Embed-Mini: Binding Modalities Without Forgetting via Dense Distillation", "published": "2026-10-01T17:49:40Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02150", "title": "From Knowledge Access to Source Learning: Developing Source-Specific Competence", "published": "2026-10-01T17:50:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02159", "title": "When Do Intrinsic Rewards Lead to Exploration?", "published": "2026-10-01T17:52:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02161", "title": "DuoMind: Enabling Distributed Multi-Robot Coordination with Semantic Communication", "published": "2026-10-01T17:53:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02163", "title": "AutoCompact: Learning When to Compact Context in Long-Horizon Coding Agents", "published": "2026-10-01T17:54:34Z", "status_at_scan": "implemented", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.02173", "title": "Every Ablation Is a Dose: Counterweights and the Semblance of Self-Repair", "published": "2026-10-01T17:57:04Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02181", "title": "OmniSeek: Native Tool Integration for Multi-turn Audio-Visual Reasoning", "published": "2026-10-01T17:58:16Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02189", "title": "Generative modeling of intrinsically disordered protein regions by reinforcing sparse autoencoder features", "published": "2026-10-01T17:59:20Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02190", "title": "Trust the Direction, Search the Step: Zero-and-First-Order Methods for LLM Fine-Tuning", "published": "2026-10-01T17:59:28Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02191", "title": "The Missing Primitive: Diagnosing and Repairing Mathematical Reasoning in Large Language Models", "published": "2026-10-01T17:59:32Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02196", "title": "InterEvolve: Test-Time Evolution of Reward Programs for Humanoid Loco-Manipulation", "published": "2026-10-01T17:59:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02199", "title": "TACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning", "published": "2026-10-01T17:59:42Z", "status_at_scan": "implemented", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02202", "title": "ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research", "published": "2026-10-01T17:59:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02204", "title": "Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents", "published": "2026-10-01T17:59:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02206", "title": "KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards", "published": "2026-10-01T17:59:55Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.02241", "title": "Hardware-Native Joint Sparse-Quantization for Trillion-Scale Mixture-of-Experts", "published": "2026-09-29T07:58:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02247", "title": "Toward Controlling Biology with Language:Offline Learning of Prompt-Conditioned Interventions for Cells, Organoids, and Biobots", "published": "2026-09-30T16:49:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02267", "title": "Fast Models, Slow Evidence: A Paired and Self-Audited Evaluation of System-1 Decision Models for LLM Agent Harnesses", "published": "2026-10-01T05:57:14Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02270", "title": "Reliability Stress Tests and Decision-Time Routing for Chest X-ray Vision-Language Models", "published": "2026-10-01T06:23:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02284", "title": "Effects of interpulse-interval variation on deep-learning classification of bat vocalizations", "published": "2026-10-01T14:01:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02298", "title": "EditHero: A Benchmark for Long-Horizon Part-Level 3D Editing and Vibe Modeling", "published": "2026-10-01T17:50:53Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02304", "title": "SimuVerity: Benchmarking Agents for Engineering-Grade Simulink Model Generation", "published": "2026-10-01T17:58:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02320", "title": "DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents", "published": "2026-10-01T18:00:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02323", "title": "World-Calibrated Proposal-to-Action Flow for Vision-Language-Action Models", "published": "2026-10-01T18:00:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02324", "title": "Slow-Fast Multi-Teacher On-Policy Distillation for Capability Preservation", "published": "2026-10-01T18:00:27Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02330", "title": "Choosing Before Acting: Comparative Value Estimation for Long-Horizon Tool-Use Agents", "published": "2026-10-01T18:04:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02344", "title": "Drive vs. Decay: On the Training Dynamics of Joint-Embedding Predictive Architectures", "published": "2026-10-01T18:16:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02349", "title": "MIRROR: Multipath Quorum Integrity for LLM Multi-Agent Communication", "published": "2026-10-01T18:22:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02351", "title": "DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents", "published": "2026-10-01T18:25:07Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02359", "title": "Lexicographic Multi-Objective On-Policy Distillation", "published": "2026-10-01T18:35:35Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02361", "title": "SEDIMA: Cross-Run Hierarchical Insight Memory for Evolutionary Search Agents", "published": "2026-10-01T18:38:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02369", "title": "Automating the Application of HCI Principles: Skills for On-Demand UI Construction, the Human-AI Space to Think, and the Future of HCI", "published": "2026-10-01T18:45:43Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02375", "title": "EviDent-CBCT: Evidence-Bottlenecked Report Generation from Dental CBCT under Non-Exhaustive Report Supervision", "published": "2026-10-01T18:54:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02376", "title": "Coco: An Agentic Copilot for the Hardware--Software Co-Design Lifecycle", "published": "2026-10-01T18:55:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02386", "title": "Social bot detection in the age of ChatGPT: Challenges and opportunities", "published": "2026-10-01T19:08:58Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02388", "title": "Octrees as an Explicit 3D Language", "published": "2026-10-01T19:10:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02396", "title": "Inherit-MAS: Test-Time Evolution of Multi-Agent Systems through Workflow and Execution Inheritance", "published": "2026-10-01T19:23:32Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02404", "title": "Trained Agentic Context Management", "published": "2026-10-01T19:32:12Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02425", "title": "Finding the Move Is Not Winning the Game: XiangqiBench for Closed-Loop Evaluation of LLM Agents", "published": "2026-10-01T19:48:08Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02432", "title": "Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations", "published": "2026-10-01T19:57:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02442", "title": "AI-driven Thermal-aware Data Center Capacity Planning", "published": "2026-10-01T20:09:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02444", "title": "Counterexample Generation via Per-Theorem Symbolic Verifiers: When Imitation Hurts and Reinforcement Repairs", "published": "2026-10-01T20:13:14Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02451", "title": "A Simulation-Grounded Agentic VLM Framework for Wildfire Monitoring and Reporting", "published": "2026-10-01T20:22:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02452", "title": "Reinforcement Learning Techniques for the Optimization of Target Polarization in Nuclear Physics Scattering Experiments", "published": "2026-10-01T20:22:54Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02460", "title": "CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarking", "published": "2026-10-01T20:34:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02462", "title": "Capability Scaling-Down Laws for LLM Compression", "published": "2026-10-01T20:35:12Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02466", "title": "SD-DPC: Sparse Dictionary Differentiable Predictive Control", "published": "2026-10-01T20:40:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02472", "title": "APDMem: Agent-Controlled Progressive Disclosure for Query-Adaptive Long-Term Memory", "published": "2026-10-01T20:51:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02478", "title": "Tropical Reinforcement Learning", "published": "2026-10-01T20:56:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02480", "title": "MEA: A Reward-Driven Multi-Agent System for Faithful Model Explanations", "published": "2026-10-01T20:57:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02486", "title": "From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders", "published": "2026-10-01T21:03:42Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02497", "title": "LiteEMG-FM: An Efficient and Deployable Foundation Model for Robust EMG Sensing", "published": "2026-10-01T21:19:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02507", "title": "MeshQuery: Agentic Seam Planning for UV Parametrization", "published": "2026-10-01T21:30:36Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02513", "title": "From Fragments to Global Maps: Learning Vectorized Map Aggregation with Large Language Models", "published": "2026-10-01T21:35:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02521", "title": "Spatial Memory Intelligence: Endowing World Models with Understanding-Driven Long-Term Memory", "published": "2026-10-01T21:50:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02523", "title": "Hypothesis-guided discovery of cognitive algorithms via program refinement", "published": "2026-10-01T21:53:21Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02525", "title": "Learning What to Investigate Next: Meta-Reasoning for Long-Horizon Research Agents", "published": "2026-10-01T21:54:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02542", "title": "How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling", "published": "2026-10-01T22:30:08Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02549", "title": "Evaluating Multi-Dimensional Generalization of Large Language Models in Temporal Extraction Tasks", "published": "2026-10-01T22:35:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02554", "title": "Test-time Multi-agent Coordination by Decomposed Value Gradient Flow", "published": "2026-10-01T22:38:27Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.02569", "title": "Pincer: Resource Authorization for Agents using a Digital Twin", "published": "2026-10-01T23:03:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02588", "title": "Open-Endedness Bench: Measuring Epistemic Process from Agent Records", "published": "2026-10-01T23:34:37Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02593", "title": "Fisher-Guided Submodular Data Selection for Continual Pre-Training of Large Language Models", "published": "2026-10-01T23:44:09Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.02597", "title": "GRAFT: Growing Agglomerative Foundation Models via Continual Teacher Distillation", "published": "2026-10-01T23:49:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02600", "title": "When History Misleads: Asymmetric Margin Supervision for Instruction-Guided LLM Generative Recommendation", "published": "2026-10-01T23:53:42Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.02610", "title": "Seer: Maximum Likelihood Regression for Learning-Speed Curves", "published": "2026-10-02T00:09:10Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.02611", "title": "Scale-Recursive Rectified Flows for Few-Step Precipitation Ensembles", "published": "2026-10-02T00:11:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02614", "title": "What Is Lost in Post-Training? Default Collapse and the Loss of In-Context Steerability Across Diverse Perspectives", "published": "2026-10-02T00:14:40Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02616", "title": "VERSE: Verified Self-Evolving Optimizer for Agent Harnesses", "published": "2026-10-02T00:16:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02617", "title": "WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness", "published": "2026-10-02T00:17:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02626", "title": "Imagine the Future, Internalize the Gist: Efficient VLA Reasoning via Internalized Spatiotemporal Imagination", "published": "2026-10-02T00:31:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02638", "title": "Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript", "published": "2026-10-02T00:54:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02654", "title": "Coherence-Driven Belief Formation and Population Dynamics of Contagion in LLM Agents", "published": "2026-10-02T01:20:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02657", "title": "Context-Tower Conversion Preserves Generation While Freezing Retains Knowledge: Low-Budget AR-to-Diffusion Conversion of MoE LLMs", "published": "2026-10-02T01:26:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02659", "title": "Distributed Learning with Selective State Space Models: Architecture-Aware Convergence Analysis", "published": "2026-10-02T01:28:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02666", "title": "CHASE-VLA: Post-Training Quantization Framework for Vision-Language-Action Models with Chunk-Aware Scale Estimation", "published": "2026-10-02T01:34:55Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02670", "title": "LEAP: Learning Efficient Action Proposals For LLM Agents", "published": "2026-10-02T01:44:03Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02683", "title": "RAOA: Alternating-Operator Neural Computation with Programmable Radio Propagation", "published": "2026-10-02T02:04:34Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02687", "title": "Decoupling Memory from Context: Structured Memory for Token-Efficient Test-Time Continual Learning", "published": "2026-10-02T02:07:32Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02695", "title": "Test-time Calibration Learning for Large Language Model Reasoning", "published": "2026-10-02T02:17:17Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02697", "title": "GeoScaffold: Learning Compact Geometric Latents via Reconstruction for Efficient Vision-Language Navigation", "published": "2026-10-02T02:24:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02700", "title": "Learning from Evolving Errors: Adaptive Iterative Repair for On-Policy Distillation", "published": "2026-10-02T02:34:57Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02702", "title": "Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise", "published": "2026-10-02T02:37:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02710", "title": "Self-Supervised Scaling of Terminal Environments for Scientific Domains", "published": "2026-10-02T02:44:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02713", "title": "WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds", "published": "2026-10-02T02:48:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02718", "title": "Revisiting Visual Representation Enhancement of VLMs via Kernel Canonical Correlation Analysis", "published": "2026-10-02T02:54:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02740", "title": "Prospective Hindsight: Self-Calibrating Reinforcement Learning via Prediction-Reality Gaps", "published": "2026-10-02T03:12:00Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02744", "title": "EpiWorld: Grounding LLM Policy Agents in Epidemiological World Models", "published": "2026-10-02T03:16:28Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.02755", "title": "FiberGeoText: A Vision-Language Model for Population- Level Organization of Superficial White Matter", "published": "2026-10-02T03:31:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02766", "title": "Exact Memory-Time Optimization for Prefix-Cached Language Model Serving", "published": "2026-10-02T03:47:10Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02781", "title": "OPD Before RL: Warm-Starting Rubric-Based RL with On-Policy Distillation", "published": "2026-10-02T04:14:57Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02814", "title": "VeriPy Source-Preserving Verification and Compatibility Checking for Python Components", "published": "2026-10-02T05:02:06Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02819", "title": "Text-Centric Post-Training for Omni-Modal Reasoning", "published": "2026-10-02T05:08:25Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02828", "title": "FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training", "published": "2026-10-02T05:19:54Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02832", "title": "FastOPD: On-Policy Distillation for Lightweight VLA Deployment", "published": "2026-10-02T05:24:20Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.02835", "title": "All Work And No Play Makes Jack a Dull Boy: Understanding and Preventing Catastrophic Strategy Collapse in RLVR", "published": "2026-10-02T05:27:33Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02839", "title": "To Explore The Strange New World Beyond Data Distribution: System Behavior, Causality Tax, and Non-causal Base Model", "published": "2026-10-02T05:29:03Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02840", "title": "PointWAM: 3D World Action Modeling for Dexterous Robotic Manipulation", "published": "2026-10-02T05:31:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02841", "title": "How Robust Is Multimodal Claim Verification to LLM Rewriting?", "published": "2026-10-02T05:31:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02847", "title": "Turnover-Orthogonal Credit Assignment for Open-Team Multi-Agent Reinforcement Learning", "published": "2026-10-02T05:35:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02856", "title": "Adaptive Mutual Distillation for Balanced Multi-Task Post-Training of Large Language Models", "published": "2026-10-02T05:42:22Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02876", "title": "Seeing, Saying, but Not Using: From Reportable Spatial Facts to Usable States in Multimodal Large Language Models", "published": "2026-10-02T06:12:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02878", "title": "Evaluating VQA in Vision Language Models using Cooperative Principles", "published": "2026-10-02T06:13:15Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02880", "title": "Found but Not Read: When Extracted Text Closes the Retrieval-Reading Gap in Document Vision-Language Models", "published": "2026-10-02T06:16:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02885", "title": "PsyEvo: A Personalized Counseling Agent That Self-Evolves at Test Time", "published": "2026-10-02T06:23:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02886", "title": "Misinformation Without Triggers: From Factual Answers to Downstream Decisions", "published": "2026-10-02T06:24:54Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02887", "title": "Revealing Epistemic Uncertainty in MLLMs via Causal-Invariant Masking", "published": "2026-10-02T06:29:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02910", "title": "Frequency Is Not Sensitivity Identifying Safety-Sensitive Experts in Sparse MoE LLM", "published": "2026-10-02T07:00:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02925", "title": "Positive-Unlabeled Learning for Agent Safety False Alarm Auditing", "published": "2026-10-02T07:16:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02928", "title": "Discriminating Fixture Coverage in Agent-Infrastructure Verification Suites", "published": "2026-10-02T07:17:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02951", "title": "Dynamic Expert Pruning for Multi-Agent Systems", "published": "2026-10-02T07:45:30Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.02952", "title": "GTDD: Generative Test-Driven Development for AI Coding Agents with Adversarial Testing", "published": "2026-10-02T07:45:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02957", "title": "Understanding Trajectory Heterogeneity in Federated World Model Learning", "published": "2026-10-02T07:55:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02970", "title": "A Guideline-Augmented Multi-Agent Framework for Schema-as-Code Biomedical Named Entity Recognition", "published": "2026-10-02T08:06:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.02974", "title": "From Language Priors to Field Adaptation: Preference Learning for Traversability Estimation", "published": "2026-10-02T08:07:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.02986", "title": "OLMo-Detect: A Multi-Stage, Confounder-Controlled Benchmark for Membership Inference on Large Language Models", "published": "2026-10-02T08:22:46Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.02994", "title": "Sentry: Learning to Recover from LLM Agent Failures at Test Time", "published": "2026-10-02T08:26:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03002", "title": "Recursive Self-Improvement in Unified Multimodal Models", "published": "2026-10-02T08:34:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03010", "title": "Engineering Sustainable Agents: A Systematic Comparison of Agentic LLMs for Developer Workflows", "published": "2026-10-02T08:41:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03014", "title": "Beyond Predefined Sinks: Security-Aware Dependency Analysis for LLM Agents", "published": "2026-10-02T08:48:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03016", "title": "From Expression to Reaction: Role-aware Visual Transfer and Stimulus-guided Reasoning for Interlocutor Emotion Recognition", "published": "2026-10-02T08:50:26Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03020", "title": "DyadMem: A Long-Term Memory Benchmark of How Agents Work with Users", "published": "2026-10-02T08:57:43Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03025", "title": "Verifiable, Articulable, and Tacit Components of Preference", "published": "2026-10-02T09:01:07Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.03033", "title": "When Numbers Start Talking: Numerical Signalling and Strategic Behaviour Among LLMs", "published": "2026-10-02T09:15:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03034", "title": "Adaptive Second-Order Solvers for Fast Stochastic Diffusion Sampling", "published": "2026-10-02T09:17:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03036", "title": "WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites", "published": "2026-10-02T09:18:28Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.03039", "title": "HyperThink: Text-to-Parameter Hypernetworks for Efficient Reasoning", "published": "2026-10-02T09:19:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03055", "title": "hacktrace: behavior-supervised detection of reward hacking during code generation", "published": "2026-10-02T09:35:59Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03056", "title": "MOF-VERIFY: A Failure-Aware Agentic Harness for MOF Hypothesis Verification", "published": "2026-10-02T09:36:55Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03057", "title": "When Does Synthetic Relational Data Teach Models to Use Relations? Tracing Predictive Structure from Pretraining Data to Model Behavior", "published": "2026-10-02T09:37:48Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03063", "title": "HARPO: Hallucination-Aware Reinforcement Learning for Faithful and Creative Language Generation", "published": "2026-10-02T09:46:40Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03084", "title": "NegT2IBench: When Negation Changes the Picture. A Polarity Benchmark for Text-to-Image Models", "published": "2026-10-02T10:03:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03089", "title": "Securing Computer-Use Agents Against Branch Steering Attacks", "published": "2026-10-02T10:08:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03098", "title": "Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression", "published": "2026-10-02T10:18:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03099", "title": "Beyond Single Videos: Benchmarking and Active Evidence Seeking for E-Commerce Cross-Video Reasoning", "published": "2026-10-02T10:19:49Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03102", "title": "Ask, Relax, or Act? Evaluating Actionable Indeterminacy in LLM Preference Reasoning", "published": "2026-10-02T10:22:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03109", "title": "Emergent Structure in the Marginal Attention Space of Language Models", "published": "2026-10-02T10:27:59Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03112", "title": "Building Interpretable Feature Representations for Resume-Vacancy Matching by Distilling Production LLM Signals", "published": "2026-10-02T10:30:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03123", "title": "Foresight: planning future perception in streaming VLMs without retraining", "published": "2026-10-02T10:44:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03135", "title": "Page-EntroKV: Hardware-Aligned, Entropy-Weighted KV-Cache Eviction under Grouped-Query Attention", "published": "2026-10-02T11:01:39Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03136", "title": "Investigating the Role of Reasoning-Language Alignment in Monolingual Retrieval-Augmented Generation", "published": "2026-10-02T11:02:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03141", "title": "Behavior Pack Optimization for Video MLLM Post-Training", "published": "2026-10-02T11:06:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03153", "title": "EvoRiskBench: An Evolving Benchmark for Runtime Security Risks in Workspace Agents", "published": "2026-10-02T11:22:36Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03185", "title": "Gains and Collapse in On-Policy Distillation:A Reinforcement Learning Perspective", "published": "2026-10-02T11:57:36Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03187", "title": "Lightweight and Resource-Efficient Perception for Robotic Guide Dogs", "published": "2026-10-02T12:01:55Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03190", "title": "Not Until the Evidence Says So: Teaching LLM Investigators When to Close a Case", "published": "2026-10-02T12:06:02Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03193", "title": "Bridging Research and Practice: A Systematic Evaluation of Generalist and Dermatology-Specific Models in Clinical Skin Lesion Classification", "published": "2026-10-02T12:09:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03195", "title": "Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It", "published": "2026-10-02T12:10:03Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.03198", "title": "KV$^2$: A Self-Refining KV Cache", "published": "2026-10-02T12:11:35Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03199", "title": "Predicting and Repairing Merge Collapse in Large Language Models", "published": "2026-10-02T12:11:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03213", "title": "Toward SLM-based agentic task-tool intent matching", "published": "2026-10-02T12:31:01Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03215", "title": "StanceEval 2026: The Second Stance Detection Shared Task", "published": "2026-10-02T12:32:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03220", "title": "Evolving Hybrid Quantum-Classical Architectures for Image Classification", "published": "2026-10-02T12:37:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03223", "title": "AdaStep: Adaptive Step Credit Weighting for Agentic Reinforcement Learning", "published": "2026-10-02T12:38:53Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.03226", "title": "D2K-Bench: Can LLM Agents Turn Expert Designs into Efficient GPU Kernels?", "published": "2026-10-02T12:42:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03252", "title": "COSMI: COmpositional Synthesis of Multi-object Interactions", "published": "2026-10-02T12:59:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03289", "title": "Architecture-Dependent Fusion Pathways in MLLMs", "published": "2026-10-02T13:32:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03315", "title": "Lightweight, Rubric-Guided Trajectory Evaluation for Production AI Agents", "published": "2026-10-02T13:51:06Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03324", "title": "To Jev or Not? Evaluating the Accuracy and Efficiency of Structured Decision Models for Hate-Speech Moderation", "published": "2026-10-02T13:59:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03356", "title": "ReFract: Benchmarking Perspective Awareness in Language Model Agents with Text World Models", "published": "2026-10-02T14:21:26Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03361", "title": "Follow the Winners: Conservative Policy Improvement with the Cross-Entropy Method for Critic-Free RFT", "published": "2026-10-02T14:24:06Z", "status_at_scan": "implemented", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.03369", "title": "Mixture-of-Experts for Cryptocurrency Order Execution: Training Stability, Tail Risk, and Failure Modes", "published": "2026-10-02T14:27:26Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03389", "title": "From Patching to Pruning Visual Computation in Vision Language Models", "published": "2026-10-02T14:37:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03421", "title": "CLIMB: Confidence-Guided Complementary Evidence for Multimodal Retrieval-Augmented Generation", "published": "2026-10-02T15:09:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03445", "title": "Corrupted but Correct: Why Vision-Language Models Lie to Themselves Internally", "published": "2026-10-02T15:27:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03448", "title": "Passing the Test You Trained On: Re-evaluating Prompt-Injection Detectors for LLM Agents", "published": "2026-10-02T15:30:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03454", "title": "Measure Less, Know More: Self-Supervised Test-Time Feature Acquisition", "published": "2026-10-02T15:33:07Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03463", "title": "Generalization of Transformer-Based Neural Quantum States via In-Context Learning", "published": "2026-10-02T15:38:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03468", "title": "A Vision-Language Model (VLM)-based Pipeline for End-to-End Procedural Modeling of Field-Grown Maize from Point Clouds", "published": "2026-10-02T15:41:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03476", "title": "MobiAgent: Dual-Loop Recursive Policy Self-Improvement for Long-Horizon Mobile Manipulation", "published": "2026-10-02T15:48:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03480", "title": "Metropolis-Hastings Dominates Importance Resampling for Policy Composition", "published": "2026-10-02T15:49:38Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03502", "title": "Certified Mechanistic Edits: Behavioral Guarantees for Skill Removal and Preservation", "published": "2026-10-02T16:01:23Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03515", "title": "Learning from Repaired Reasoning: Root-Cause-Guided On-Policy Distillation", "published": "2026-10-02T16:07:39Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03516", "title": "XGenAct: Geometry-Enhanced World Action Models through Cross-Task Generation", "published": "2026-10-02T16:10:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03524", "title": "From Benchmarks to Production: A Text-to-SQL System for Complex Financial Data", "published": "2026-10-02T16:13:40Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03525", "title": "Structured Composition of Verifiable Atomic Insights for Table-to-Report Generation", "published": "2026-10-02T16:14:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03529", "title": "Divergence controls entropy in distillation", "published": "2026-10-02T16:15:53Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.03567", "title": "Writerslogic at the CLEF 2026 SimpleText Track: Multi-Candidate LLM Simplification and Stacked Complexity Spotting", "published": "2026-10-02T16:44:53Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.03574", "title": "HyperBrowseComp: A Multilingual and Multimodal Stress Test for Web-Browsing Agents", "published": "2026-10-02T16:48:49Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03585", "title": "Threat-Preserving Representation Sensitivity in Agent-Security Benchmarks", "published": "2026-10-02T16:55:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03591", "title": "HazardWeaver: Scientific Route Selection for Hazard Analysis Agents", "published": "2026-10-02T16:59:19Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03598", "title": "When a Correct Reward Is Not Enough: Diagnosing and Guiding PPO in an Analytically Solved Broker-Trader Game", "published": "2026-10-02T17:03:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03604", "title": "Mastering Atari 2600 Games with Discovered Options", "published": "2026-10-02T17:05:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03617", "title": "DEPICT: Scoring Text-to-Image Alignment by Answer Agreement", "published": "2026-10-02T17:13:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03620", "title": "UniIntervene++: An Adaptive Intervention Agent for Efficient Real-World Reinforcement Learning", "published": "2026-10-02T17:18:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03631", "title": "NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents", "published": "2026-10-02T17:23:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03632", "title": "World Embedding Benchmark", "published": "2026-10-02T17:24:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03634", "title": "Credit Where It Matters: Dependency-Aware Policy Optimization for Terminal Agents", "published": "2026-10-02T17:24:48Z", "status_at_scan": "implemented", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03665", "title": "Pivot-SD: Efficient Self-Distillation for Masked Diffusion Language Models", "published": "2026-10-02T17:37:51Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03675", "title": "FrugalEvo: Towards Cost-Aware LLM-Guided Program Evolution", "published": "2026-10-02T17:44:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03695", "title": "Language Models that Play Chess and Explain Their Moves", "published": "2026-10-02T17:54:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03702", "title": "LESSER: Post-Training Data Selection with Output-Layer Gradients", "published": "2026-10-02T17:55:42Z", "status_at_scan": "implemented", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.03713", "title": "What Should World Models Forget? Stratified Retention for Continual Adaptation", "published": "2026-10-02T17:58:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03741", "title": "Logit-Aware MIMO AirComp for Distributed Mixture-of-Experts LLM Inference over Wireless Edge Networks", "published": "2026-09-22T15:55:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03759", "title": "BridgeCast: Bridging Ocean Wave Forecasts to Reanalysis via Flow Matching with Exogenous Variables", "published": "2026-09-27T00:43:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03780", "title": "OceanMind: A multi-agent AI system for ocean diagnosis", "published": "2026-09-30T06:51:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03792", "title": "What Do Verifiable Rewards Teach Video-Language Models About Time? A Controlled Multi-Model Study", "published": "2026-09-30T21:42:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03796", "title": "ProsaBuddy: Assisting Mechanized Real-Time Schedulability Analysis with LLM-based Agents", "published": "2026-10-01T03:29:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03820", "title": "From Requirements to Attack Trees: Grounded LLM Agents for Design-Time Security Review", "published": "2026-10-02T03:52:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03830", "title": "Memory-State Critic for Asymmetric Actor-Critic with Application to Vision-Based Pursuit-Evasion", "published": "2026-10-02T11:58:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03894", "title": "Proxy Confidence: Auditing Black-Box LLM Agents with a Surrogate's Log-Probabilities", "published": "2026-10-02T18:09:00Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03902", "title": "When Evidence Changes: Evaluating Memory Repair and Re-reading in Language-Model Agents", "published": "2026-10-02T18:13:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03923", "title": "Learning Robust Personalized Prompts for LLM-Driven Sequential Recommendation", "published": "2026-10-02T18:34:26Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.03935", "title": "General Decision Models: Benchmarking and Insights Beyond Jev", "published": "2026-10-02T18:47:00Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.03938", "title": "MLLMs Fail to Refuse when Using Tools Agentically", "published": "2026-10-02T18:48:51Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03969", "title": "How Do Coding Agents Optimize Software and Report Performance Validation? A Large-Scale Empirical Study of Open-Source Pull Requests", "published": "2026-10-02T19:23:43Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.03972", "title": "Probabilistic Algorithms for Ising Machines from Optimization to Generative AI", "published": "2026-10-02T19:30:02Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.03984", "title": "Teaching Agents to Code Reliably", "published": "2026-10-02T19:47:19Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.03994", "title": "Solving VeriContest with a Lean-Backed Rust Verifier", "published": "2026-10-02T19:56:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04002", "title": "Do Language Models Need a Trainable Input Embedding Table? Fixed Minimal Token Codes at 1.7B-Class Scale", "published": "2026-10-02T20:06:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04008", "title": "SkillScriptBench: Benchmarking Self-Evolution of Executable Agent Skill Packages Beyond Markdown", "published": "2026-10-02T20:10:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04009", "title": "SUAVE: Unified Video-Action Models via Masked Diffusion", "published": "2026-10-02T20:11:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04040", "title": "Agent Policy-Value Audit: Separating Transition Composition from Event Selection in Financial LLM Agents", "published": "2026-10-02T20:46:34Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04047", "title": "Periscope: Extending Frozen Language Models Beyond Their Context Window", "published": "2026-10-02T20:52:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04053", "title": "The Cost of a Hop: Benchmarking NLIP and A2A", "published": "2026-10-02T21:04:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04083", "title": "Self-Propagating Misalignment in LLM Agents, and Why Auditing or Disabling Memory Is Not Enough", "published": "2026-10-02T21:46:13Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04104", "title": "Where Does the Semantic Gain Come From? A Reproduction and Extension of Semantic Knowledge-driven Contrastive Learning for Long-Tailed Recognition", "published": "2026-10-02T22:17:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04118", "title": "LongSocialBench: Do Long-Context LLMs Understand Online Discussion Threads?", "published": "2026-10-02T22:47:29Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04125", "title": "Representational Control over Self-Report & Behavior Coherence in LLM Risk-Taking", "published": "2026-10-02T22:54:43Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04132", "title": "PB-GRPO: Learning Socially Adaptive LLM Agents from Persona-Driven Simulation with Preference-Batched GRPO", "published": "2026-10-02T23:07:24Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04133", "title": "Auditing Pairwise Equivalence Judgments: Self-Critique Effects and Diversity Measurement in Multi-Agent Hypothesis Generation", "published": "2026-10-02T23:07:36Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04134", "title": "Agentic Resource Allocation for Batch Multi-Objective Bayesian Optimization in Autonomous Materials Discovery", "published": "2026-10-02T23:08:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04138", "title": "Dual-Scale Relational Graph Transformers for Ecosystem-Aware Fraud Detection", "published": "2026-10-02T23:21:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04139", "title": "From Sight to Foresight: Predictive Spatial Reasoning in Vision-Language Models", "published": "2026-10-02T23:27:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04140", "title": "ExpertMuon-Compass: Alignment-Guided Step Sizes for Mixture-of-Experts Training", "published": "2026-10-02T23:27:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04152", "title": "Kepler4D: Controllable Future Video Generation via 4D Scene State Evolution", "published": "2026-10-02T23:53:00Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04156", "title": "Trajectory-Derived Confidence for Reliable, Resource-Aware Clinical Text-to-SQL Agents", "published": "2026-10-02T23:57:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04158", "title": "How RL Reshapes LLM Reasoning: Transferability, Coverage, and Scaling Laws", "published": "2026-10-03T00:01:36Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04168", "title": "Agentic Cognitive Depth: Operational Criteria for Evaluating LLM Agents", "published": "2026-10-03T00:19:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04178", "title": "SHarP: Saliency-based Pruning of Agent Harnesses", "published": "2026-10-03T00:41:14Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04196", "title": "Asynchronous Is Nearly Free for Evolution Strategies on Long-Horizon Agentic Tasks", "published": "2026-10-03T01:18:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04204", "title": "Clean: Second-order LLM Training at Linear Memory Cost via Nyström Sketching", "published": "2026-10-03T01:37:36Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04225", "title": "FlashGaze: Training-Free Multi-Scale Patch Pruning For Efficient Video Understanding", "published": "2026-10-03T02:33:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04246", "title": "Benchmarking Psychological Dynamics in Generative Agents", "published": "2026-10-03T03:08:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04255", "title": "Grounded in Time: A Multi-Source Dataset and Benchmark for Temporal Grounding in Robotic Manipulation", "published": "2026-10-03T03:25:24Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04261", "title": "Playing social deduction games with reinforcement fine-tuned large language models", "published": "2026-10-03T03:37:15Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04263", "title": "Multimodal Dual-Encoder Retrieval for Automated ICD Coding", "published": "2026-10-03T03:39:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04269", "title": "Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure Experience", "published": "2026-10-03T03:45:44Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04272", "title": "Rethinking Self-Distillation for Multi-Teacher Capability Merging", "published": "2026-10-03T03:57:33Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04278", "title": "Do RUL explanations hold up? Faithfulness and stability of attributions on C-MAPSS", "published": "2026-10-03T04:10:39Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04292", "title": "LMBuild: Evaluating LLM Agents for Generating Buildable and Functional Structures", "published": "2026-10-03T04:55:57Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04299", "title": "Questioning the Questions: Sustaining Self-Evolution in Reasoning Models", "published": "2026-10-03T05:24:47Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04302", "title": "From Valid to Useful: Post-Verification Acquisition for Recursive Self-Improving Recommendation", "published": "2026-10-03T05:31:09Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.04303", "title": "What to Preserve in Recursive Computation: A Local Predictive Sufficiency Principle", "published": "2026-10-03T05:33:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04318", "title": "Rethinking Long-Video Efficiency: A Joint Allocation Perspective on Frames, Pixels, and Front-End Latency", "published": "2026-10-03T06:13:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04328", "title": "Bidirectional Preference Synthesis: Learning Prompt-Conditioned Preferences from Boundary Failures", "published": "2026-10-03T06:48:35Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04344", "title": "Hierarchical Credit Assignment for RLVR on Fused Gromov-Wasserstein Geometry", "published": "2026-10-03T07:11:02Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04366", "title": "Human Behavior-Informed Crash Scenario Generation with Real-World Crash Priors for Autonomous Vehicle Safety Evaluation", "published": "2026-10-03T08:04:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04371", "title": "Functionally Equivalent or Not? Graph-Grounded Differential Surrogate Execution for Code Equivalence", "published": "2026-10-03T08:19:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04375", "title": "Do Tool Calls Execute as Intended? Measuring and Repairing Intent-Execution Correspondence in LLM Agents", "published": "2026-10-03T08:38:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04378", "title": "COPEX: Benchmarking LLM Robustness to Adversarial Context Across Model Context Protocol Layers", "published": "2026-10-03T08:44:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04379", "title": "AgentPersonaBench: Benchmarking Persona-Driven User Simulation", "published": "2026-10-03T08:45:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04381", "title": "Beyond Plausibility: Verifiable Fine-Grained Image Editing on Structured Assets", "published": "2026-10-03T08:48:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04403", "title": "Saying, Not Knowing: Aggressively GGUF-Quantized Small Language Models Still Write Rare Words They Can No Longer Define", "published": "2026-10-03T09:50:16Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.04409", "title": "Understanding and Mitigating Hallucination Escape in Tool-Using LLM Agents", "published": "2026-10-03T09:57:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04412", "title": "Large Language Models and Augmented Democracy", "published": "2026-10-03T10:02:14Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04413", "title": "Specific Algorithmic Interpretability of Neural Networks: A Case Study on Textures", "published": "2026-10-03T10:03:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04418", "title": "CORE-RL: Confidence-Oriented Reliability Evaluation of Black-Box Reinforcement Learning Policies", "published": "2026-10-03T10:15:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04424", "title": "On the Trade-off Between Information Loss and Generalization in Sparse Attention", "published": "2026-10-03T10:27:29Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04425", "title": "AgroGround: Multi-Granularity Grounded Recognition in Agriculture", "published": "2026-10-03T10:27:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04429", "title": "Asking Earns Nothing: Scoring the Decision to Act in BFCL Multi-Turn", "published": "2026-10-03T10:42:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04432", "title": "Video2World: Benchmarking Coding Agents for Interactive World Modeling from Embodied Videos", "published": "2026-10-03T10:46:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04433", "title": "What Does a Harness Buy? Tokens, Mostly", "published": "2026-10-03T10:47:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04441", "title": "LocusRL: Diagnosing LLM Reward and Policy Interventions in Competitive Games", "published": "2026-10-03T11:03:16Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04488", "title": "Can LLM Agents Automate Reinforcement Learning for Text-to-Speech?", "published": "2026-10-03T12:44:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04494", "title": "DreamTest: World-Model Surrogates for Search-Based Testing of Deep Reinforcement Learning Agents", "published": "2026-10-03T12:51:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04499", "title": "Homogeneous Semantic Alignment and Hierarchical Expert Routing for Radiology Report Generation", "published": "2026-10-03T12:57:35Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04502", "title": "Localization Lens for Improving Medical Vision-Language Models", "published": "2026-10-03T13:04:15Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04504", "title": "The Same Zero: Why Identical ASR Can Imply Different Guarantees in LLM-Agent Security", "published": "2026-10-03T13:09:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04506", "title": "EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning", "published": "2026-10-03T13:09:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04507", "title": "LoRA's Second Descent Extends Beyond Parameter Parity", "published": "2026-10-03T13:10:37Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.04512", "title": "Correctness Is a Direction: Geometric Answer Selection in Language Models", "published": "2026-10-03T13:21:04Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04517", "title": "EvoCast: Reliable Autonomous Research Agents for Iterative Forecasting Architecture Evolution", "published": "2026-10-03T13:30:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04518", "title": "Length Generalization Needs Proper Regularization", "published": "2026-10-03T13:31:00Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04528", "title": "Quantifying Collusion Among Autonomous LLM Agents: A Statistical Analysis of the Collusion Wiki Incident", "published": "2026-10-03T13:46:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04541", "title": "Autonomous Structuring of Radiology Reports Across Modalities at Archive Scale Using an Open-Weight Large Language Model", "published": "2026-10-03T14:15:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04544", "title": "Label Agreement Does Not Measure Authorization", "published": "2026-10-03T14:17:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04555", "title": "$\\mathrm{TRIZ}^{a}$: Guiding Agent Evolution from Pattern Recognition to Solution Invention", "published": "2026-10-03T14:44:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04569", "title": "From Transformers to Weighted Automata: Towards the Verification of Large Language Models", "published": "2026-10-03T15:04:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04575", "title": "From Probe Scores to Alarm Policies: Operational Validity of Activation Monitors for Language-Model Agents", "published": "2026-10-03T15:11:21Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04580", "title": "Strong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs", "published": "2026-10-03T15:15:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04589", "title": "StegoMemory: Agentic Memory Acts as Covert Steganographic Channel", "published": "2026-10-03T15:23:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04596", "title": "DiffGate: Difficulty-Gated Teacher Guidance for On-Policy Distillation", "published": "2026-10-03T15:33:21Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04604", "title": "Anticipating the Consequences of Curriculum Decisions with Large Language Models", "published": "2026-10-03T15:45:17Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04607", "title": "ForeAct3D: Policy-Grounded Future World Modeling for VLA Policies", "published": "2026-10-03T15:49:04Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04611", "title": "Practical Runtime Enforcement of First-Order Temporal Requirements", "published": "2026-10-03T15:52:08Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04616", "title": "PerturBot: Breaking Shortcut Priors in Vision-Language-Action Models with Perturbative Training", "published": "2026-10-03T15:57:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04636", "title": "Does Neural Complexity Improve Health Misinformation Detection? A Leakage-Controlled Cross-Corpus Benchmark", "published": "2026-10-03T16:26:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04639", "title": "When Talk Isn't Code: Comparing LLM Agents That Simulate and Develop Software", "published": "2026-10-03T16:29:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04641", "title": "Back to the Future: Regressing Readability Features from LLMs", "published": "2026-10-03T16:31:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04646", "title": "SEIS: Self-Evolving Inference Systems", "published": "2026-10-03T16:36:48Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04651", "title": "GS-Codec: A Gaussian-Splatting Bottleneck for Neural Audio Coding", "published": "2026-10-03T16:49:17Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.04658", "title": "RETRACE: From Entangled Repair Histories to Reusable Experience for CI Repair", "published": "2026-10-03T17:08:17Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04672", "title": "MASBench: Benchmarking LLM-based Multi-Agent Collaboration under Partial Observability", "published": "2026-10-03T17:40:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04687", "title": "Understanding Errors in LLM-Based Question Answering over Imperfect Tables", "published": "2026-10-03T18:03:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04699", "title": "Not Self-Decidable: LLMs Cannot Draw the Boundary of What an Agent Verifier Can Check", "published": "2026-10-03T18:25:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04716", "title": "Towards Automatically Pruning Logging Code with Coding Agents: How Far Are We?", "published": "2026-10-03T19:11:19Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04721", "title": "Knossos and Ariadne: Benchmarking and Learning Complete Diagram Topology Extraction with Vision-Language Models", "published": "2026-10-03T19:20:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04740", "title": "Toward a Locally Deployable Agentic Co-Scientist: Small-Model Planning for Early-Stage Drug Discovery", "published": "2026-10-03T20:09:34Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04753", "title": "More Value per Key: Asymmetric Sparse Attention for Faster LLM Decoding", "published": "2026-10-03T20:38:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04777", "title": "ARISE: Adaptive Agentic Reasoning with Image-grounded Self-Evaluation for Interpretable IBD Assessment", "published": "2026-10-03T21:35:55Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.04779", "title": "Asymmetric Repository Lineage Modeling and Verifier-Guided Coordination in Concurrent AI Coding Agents", "published": "2026-10-03T21:37:15Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04792", "title": "Do More Modalities Always Help? A Geometric Perspective on Missing-Modality Robustness", "published": "2026-10-03T22:34:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04794", "title": "Knowing the Store: What a Memory Backend Must Write Down Before an Agent Can Read It", "published": "2026-10-03T22:39:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04802", "title": "Multi-Agent Spectrum Sharing", "published": "2026-10-03T23:04:54Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04805", "title": "ExStereo: Lifting 2D Vision-Language-Action Models to 3D with Explicit Stereo Representations", "published": "2026-10-03T23:21:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04824", "title": "Agent Behavior as Code: Efficient and Robust LLM Agents with Programmatic Specifications", "published": "2026-10-04T00:14:33Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.04832", "title": "Agent Skill Evolution: How Revisions Affect Coding Agents", "published": "2026-10-04T00:33:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04845", "title": "Which Preferences to Train On? End-to-End Multi-Objective Alignment with an Adversarial Preference Distribution", "published": "2026-10-04T01:09:45Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04851", "title": "CURIO: Curiosity-Driven Test-Time Learning for Open-Ended Discovery", "published": "2026-10-04T01:30:47Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04862", "title": "GitSwarm: Decentralized Compounding Inference", "published": "2026-10-04T02:03:24Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04867", "title": "Greedy Local Learning for Language Model Pretraining: Gaps and Objective Design", "published": "2026-10-04T02:10:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04868", "title": "From Memory to Guide: Spatio-Temporal Composer for Procedural Coding Memory", "published": "2026-10-04T02:15:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04878", "title": "Bounded Reasoning: Cognitive Hierarchy in Human-versus-AI Cyber Defense", "published": "2026-10-04T02:29:26Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04882", "title": "ResOPD: Tail Residualization for Sparse On-Policy Distillation", "published": "2026-10-04T02:43:39Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04888", "title": "No Hindsight for LLM Fact-Checkers: Measuring Leakage Channels in Misinformation Detection", "published": "2026-10-04T02:53:23Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04889", "title": "ForkPilot: Self-Evolving Policy for Retrospective Search in Long-Horizon Agents", "published": "2026-10-04T02:53:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04893", "title": "SFlexRCA: Lightweight, Scalable, and Flexible Root Cause Analysis for IIoT Edge Systems", "published": "2026-10-04T03:01:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04899", "title": "Rewrite What Matters: Adaptive Multilingual Query Rewriting for Reasoning via Agentic Reinforcement Learning", "published": "2026-10-04T03:19:46Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.04914", "title": "Monitorability Disposition in Large Reasoning Models", "published": "2026-10-04T03:49:36Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04915", "title": "Are We Measuring Scientific Intelligence? Rethinking the Evaluation of AI Scientists", "published": "2026-10-04T03:50:36Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04916", "title": "PreAct-Nav: Agentic Reasoning Before Action for Urban Navigation", "published": "2026-10-04T03:50:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04918", "title": "Residual Visual Credit Optimization: Conserved Evidence Routing for Multimodal Reinforcement Learning", "published": "2026-10-04T03:56:38Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.04921", "title": "Complex Agents, Shallow Tests: Demystifying and Enhancing Test Adequacy of Agent Harness in the Wild", "published": "2026-10-04T04:03:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04927", "title": "Assembling Insights for Agentic Machine Learning Engineering Systems", "published": "2026-10-04T04:10:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04940", "title": "Software World Models: From Consequence Prediction to Decision Value", "published": "2026-10-04T04:34:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04950", "title": "How Should Teachers Be Prepared? RL on Student-Induced States for On-Policy Distillation", "published": "2026-10-04T04:49:24Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.04957", "title": "Trinity: One Differentiable Physics for Training, Refining and Scoring Generative Floorplanners", "published": "2026-10-04T05:07:16Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04961", "title": "Building LLM Agent Systems the Deep Learning Way: From Modular Design to Architecture Search", "published": "2026-10-04T05:11:20Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.04973", "title": "TrajLong: Co-Designing Agentic and Long-Context Supervision for Mid-Training", "published": "2026-10-04T05:42:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04976", "title": "Priced Guidance: Can Language Models Generate Future Research Ideas?", "published": "2026-10-04T05:45:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.04977", "title": "Do LLMs Understand Sequential Structure? A Controlled Study of Inference and Generation", "published": "2026-10-04T05:45:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.04991", "title": "CIPO: Counterfactual Imagination Policy Optimization for Adaptive Tool Granularity Selection", "published": "2026-10-04T06:20:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05009", "title": "Rethinking Tool Design for Agentic RCA: A Controlled Empirical Study", "published": "2026-10-04T07:04:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05023", "title": "Look Where You Say You're Looking: Self-Grounded Attention for Visual Reasoning", "published": "2026-10-04T07:37:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05024", "title": "LightVLN: Efficient Aerial Vision-and-Language Navigation with Compact Memory and History-Guided Local Aggregation", "published": "2026-10-04T07:37:44Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05025", "title": "Triggering Generalist Reasoning via Predictive Uncertainty for Dual-System VLA", "published": "2026-10-04T07:39:20Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05026", "title": "GeoBridge-VLA: Geometry-Aware Residual Adaptation for Vision-Language-Action Models", "published": "2026-10-04T07:42:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05029", "title": "SPACE-CLIPv2: Decoding Local Geometry from Frozen CLIP for Monocular Depth Estimation", "published": "2026-10-04T07:50:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05030", "title": "EVISKILL: Grounding Skill Evolution in Replayable Evidence", "published": "2026-10-04T07:50:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05036", "title": "Characterizing Security Effects of OSS Vulnerabilities in Agent Systems", "published": "2026-10-04T08:06:34Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05037", "title": "SparseCraft: Agentic Hardware-Software Co-Optimization for Sparse Computing", "published": "2026-10-04T08:06:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05039", "title": "Causal Improvement Graph for Agentic Harness Optimization", "published": "2026-10-04T08:15:37Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05041", "title": "Communication Shapes Collective Inference in Self-Adapting LLM Societies: Evidence from Mafia", "published": "2026-10-04T08:19:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05047", "title": "Beyond Task Completion: Measuring Interaction Cost in Terminal User Interfaces", "published": "2026-10-04T08:30:26Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05066", "title": "Salvation Lies Within: Eliciting Inherent Style Transfer in Step-Distilled Diffusion Models", "published": "2026-10-04T09:06:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05097", "title": "ReMAP: Restoring the Perceptual Cycle with Reasoning-Time Latent Visual Memory", "published": "2026-10-04T10:00:49Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05099", "title": "Small Agents with Semantic Search: Efficient Multilingual Code Localization", "published": "2026-10-04T10:13:41Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.05100", "title": "METRO: Metric-Enhanced Token Routing Operator", "published": "2026-10-04T10:14:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05103", "title": "VulValidate: Auditing Function-Level Vulnerability Labels with Executable Evidence", "published": "2026-10-04T10:21:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05107", "title": "SearchJev: A Fast and Calibrated System-1 Model for Search Agents", "published": "2026-10-04T10:30:40Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05111", "title": "InstMoE: Adaptive Multimodal Routing with Specialized Experts", "published": "2026-10-04T10:40:07Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05114", "title": "Belief-Trajectory Energy: Measuring the Path to a Prediction", "published": "2026-10-04T10:44:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05119", "title": "LexiHorizon: Stabilizing Reinforcement Learning for Long-Horizon Deep Search", "published": "2026-10-04T10:53:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05124", "title": "Memory Canonicalization: A Framework and Benchmark for Cross-Model Drift in Persistent LLM Memory", "published": "2026-10-04T11:09:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05125", "title": "Bayesian Entropy-based Reordering for Calibrated Diffusion Language Models", "published": "2026-10-04T11:13:21Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05126", "title": "Selecting Repetition Counts Across Model Scales in Data-Constrained Pretraining", "published": "2026-10-04T11:16:04Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05129", "title": "Representation--Behavior Alignment for Explainable Weakly-Supervised Video Anomaly Detection", "published": "2026-10-04T11:21:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05131", "title": "RoMod: Temporal Routing Modulation via Mixture-of-Experts for Video Anomaly Detection", "published": "2026-10-04T11:23:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05132", "title": "SpecAgent: Empowering Program Verification with Agentic Synthesis of Formal Program Specifications", "published": "2026-10-04T11:24:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05139", "title": "Hidden in the Comments: A Context-Injection Attack Surface in Code LLMs", "published": "2026-10-04T11:39:57Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05162", "title": "Memadapter: Counterfactual Adaptation Against Memory-induced Sycophancy", "published": "2026-10-04T12:12:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05176", "title": "AECG: Asymmetric Experience Consolidation and Governance In Multi-Agent Systems", "published": "2026-10-04T12:38:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05185", "title": "Recurrent Latent Visual Search for GUI Grounding", "published": "2026-10-04T12:49:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05190", "title": "FORGE: Verification-Gated Behavioral Repair for Generative Language Models", "published": "2026-10-04T12:55:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05200", "title": "Learning without Overwriting: A Theory of Self-Distillation and Supervised Fine-Tuning in Continual Reasoning", "published": "2026-10-04T13:13:18Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05219", "title": "Safe Context Switching for Agents in the Wild: Mitigating Subspace Interference via Orthogonal Adaptation", "published": "2026-10-04T13:39:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05220", "title": "Ranking Bandits for Carousel Interfaces with Observable Browsing Depth", "published": "2026-10-04T13:39:38Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05223", "title": "Look Before You Leap: Thermodynamic Arbitration of Parametric and Non-Parametric Knowledge in LLM Agents via Self-Regulating Memory Architectures", "published": "2026-10-04T13:43:24Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.05241", "title": "StateWise: Diagnosing and Repairing Persistent Operational State Before Agent Actions", "published": "2026-10-04T14:12:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05247", "title": "templar: agentic induction and evolution of standardized radiology reporting templates from large-scale clinical corpora", "published": "2026-10-04T14:25:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05249", "title": "ArticuTable: Generating Instance-Level Interactive Rigid-Articulated 3D Tabletop Scenes from a Single Image", "published": "2026-10-04T14:26:34Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05265", "title": "Loopy: Low-Bit Quantization Framework for Looped Language Models", "published": "2026-10-04T14:35:37Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05273", "title": "When and What to Prune? Stage-Aware Visual Token Pruning for Efficient VLA", "published": "2026-10-04T14:47:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05274", "title": "Answer with Evidence: Consistency-Aware Grounded Visual Question Answering for Roadside Traffic Scenes", "published": "2026-10-04T14:48:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05282", "title": "Red-TTT: Test-Time Training for Automated Jailbreaking Large Language Models", "published": "2026-10-04T15:03:08Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05284", "title": "Fusion is the New Mutation: Bandit-Guided Evolution on Workflow Graphs", "published": "2026-10-04T15:04:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05292", "title": "Erased, Rerouted, or Rescaled? Post-Training and the Causal Quotient of a Language Model's Belief State", "published": "2026-10-04T15:11:16Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05295", "title": "Readable Before Actionable: Causal Tracing of Indirect Prompt Injection", "published": "2026-10-04T15:16:13Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05300", "title": "MESH-Harness: Self-Improving Agent Harnesses via Bandit-Guided Compositional Evolution", "published": "2026-10-04T15:22:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05303", "title": "ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience", "published": "2026-10-04T15:27:20Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05305", "title": "Characterizing Parallelism Strategies in LLM Inference: Fundamental Compute-Communication Trade-offs", "published": "2026-10-04T15:28:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05306", "title": "Kinematics-Centric Continuous Sign Language Retrieval with Gloss-Guided Boundary-Aware Alignment", "published": "2026-10-04T15:28:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05308", "title": "RubricArmor: Adversarial Evolution Improves LLM-Based Rubric Generation", "published": "2026-10-04T15:30:32Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05318", "title": "Robust Parameter-Efficient LLM Adaptation on Analog Hardware", "published": "2026-10-04T15:39:20Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05319", "title": "Understanding the Hierarchical Structure and Functional Landscape of the Model Context Protocol Ecosystem", "published": "2026-10-04T15:39:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05329", "title": "Understanding the Weight Averaging Mechanism in LLM Training for Post-Training Quantization", "published": "2026-10-04T15:53:28Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05334", "title": "AgentDiscover: Autonomous Discovery with Minimal Search Scaffolding", "published": "2026-10-04T15:55:47Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.05342", "title": "IRSTD-Agent: Agentic Infrared Small Target Detection via Zoom-Guided Interaction Learning", "published": "2026-10-04T16:10:08Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.05367", "title": "AIProver: Agentic Auto-Formalization of Mathematical Research via Certificate-Driven Evolving Harness", "published": "2026-10-04T16:48:12Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.05369", "title": "AutoDP-LLM: Automating Data Pre-processing for Intrusion Detection Systems using Large Language Models", "published": "2026-10-04T16:50:21Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05373", "title": "Towards Unbiased On-Policy Distillation for Block Diffusion Language Models", "published": "2026-10-04T17:00:16Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05380", "title": "VHDL-REPOBENCH: A Repository-Level Benchmark for Evaluating Large Language Models on VHDL Design Generation", "published": "2026-10-04T17:12:34Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05383", "title": "Sibyl: An Efficient Small-large Model Collaboration Framework for Long-horizon Tasks", "published": "2026-10-04T17:14:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05387", "title": "GNN-CB: A Graph Neural Network Competition Benchmark for Human and LLM Evaluation", "published": "2026-10-04T17:16:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05398", "title": "MMPostTrainBench: Benchmarking Autonomous Research for Multimodal Post-Training", "published": "2026-10-04T17:32:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05399", "title": "Adaptive Code Revision Attacks on AI Pull Request Reviewers", "published": "2026-10-04T17:34:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05400", "title": "CASE: Cost-Aware Stopping for Efficient Long-Video Agents", "published": "2026-10-04T17:35:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05402", "title": "Task Vector Descent: Learning from Non-IID Batches", "published": "2026-10-04T17:37:49Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05413", "title": "A Strong Baseline for Evaluating Vision Encoders in Multimodal Large Language Models", "published": "2026-10-04T17:54:09Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05417", "title": "Render to Reason: Novel-View Semantic Prediction Improves Spatial Understanding in VLMs", "published": "2026-10-04T17:58:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05418", "title": "EvoMem-VLA: State-Evolution Memory for Long-Horizon Robot Manipulation", "published": "2026-10-04T17:59:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05425", "title": "Unmentioned Checklist Findings Change How Reinforcement Learning Appears to Improve Chest Radiograph Report Checking", "published": "2026-10-04T18:06:43Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05446", "title": "Hierarchical Time-aware Bootstrapping for Off-Policy Subgoal Value Learning", "published": "2026-10-04T18:33:23Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.05461", "title": "AI Safety via Debate is Compromised by Cognitive Biases", "published": "2026-10-04T19:11:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05463", "title": "Human-Like Attention? A Psychophysical Comparison of Visual Search in Humans and MLLMs", "published": "2026-10-04T19:11:35Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05473", "title": "Hierarchical Reinforcement Learning with Stable Temporal Abstraction for Language Model Agents", "published": "2026-10-04T19:25:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05481", "title": "CodeForge-MA: Execution-Verified Multi-Agent Learning with Language-Conditioned LoRA for Multilingual Code Generation", "published": "2026-10-04T19:40:17Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.05484", "title": "Universal Test-Time Training", "published": "2026-10-04T19:43:36Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05501", "title": "Logic-Logit: A Logic-Based Approach to Choice Modeling", "published": "2026-10-04T20:14:20Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05532", "title": "DelegationBench: Measuring When AI Agents Should Ask Before Acting", "published": "2026-10-04T20:54:35Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05538", "title": "LiFT: Loop Flow Transformers", "published": "2026-10-04T21:11:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05540", "title": "SALUS: Automated Auditing of NL-to-SQL Benchmarks through Weak Supervision of Multi-Agent Output", "published": "2026-10-04T21:12:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05559", "title": "Cut Binary Cross Entropy: Efficient Large-Vocabulary Loss and Gradient Kernels for Sequential Recommendation", "published": "2026-10-04T21:47:02Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05563", "title": "G-CARB: Graph-Localized Conformal Agent Risk Budget for Compositional Harm", "published": "2026-10-04T21:49:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05569", "title": "Factoriax: A GPU-Accelerated Factorio-Style Simulator for Reinforcement Learning", "published": "2026-10-04T22:02:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05572", "title": "Disentangling Task Difficulty from Run-Level Failure in Agent Failure Prediction", "published": "2026-10-04T22:17:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05586", "title": "AgentDoxx: Agentic Re-identification of Anonymized Text with Web Search", "published": "2026-10-04T22:41:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05591", "title": "What Is a Repeated Token Worth? The Scaling Geometry of Multi-Epoch Pretraining", "published": "2026-10-04T22:55:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05600", "title": "An LLM-in-the-loop RL Framework for Bioinformatics Feature Selection", "published": "2026-10-04T23:05:21Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.05615", "title": "DREAM: Dynamic Resolution Assignment For Multimodal Multi-agent Debate", "published": "2026-10-04T23:23:02Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.05622", "title": "UndoBench: Separating Task Competence from Recovery Capability in Tool-Using AI Agents", "published": "2026-10-04T23:36:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05630", "title": "Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection", "published": "2026-10-04T23:46:35Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05637", "title": "Visual Grounding Safety in Vision-Language Models", "published": "2026-10-05T00:14:23Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05664", "title": "StageVLN: Spatial and Trajectory Auxiliary Guidance for Efficient Vision-Language Navigation", "published": "2026-10-05T01:20:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05670", "title": "Generate What You Can Trust: Content Credibility in Generative Recommenders", "published": "2026-10-05T01:33:29Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05681", "title": "Automatic Speech Recognition for Low-Resource Sinhala: A Critical Review of Methods, Challenges, and Future Directions", "published": "2026-10-05T01:49:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05685", "title": "Spend Bytes on Breadth: Precision-Count Trade-offs for Decode-Time KV Compression in Long Chain-of-Thought Reasoning", "published": "2026-10-05T01:56:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05690", "title": "Errors of LLM-Assisted Literature Retrieval in Environmental Science: A Comparison Study of Abstract versus Full-text Based Prompts", "published": "2026-10-05T02:04:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05697", "title": "From Token-Max to Outcome-Max: How You Use AI Determines Its Productivity", "published": "2026-10-05T02:16:26Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05709", "title": "Nexus: An Execution Fabric for AI Agents Across Cloud, Edge, and Devices", "published": "2026-10-05T02:47:01Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05715", "title": "Rotated, but How Far? Diagnosing and Improving Object-Rotation Reasoning in VLMs", "published": "2026-10-05T02:55:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05744", "title": "CIPHER-MoE: Balancing Efficiency and Routing Fidelity in Trillion-Scale MoE Training", "published": "2026-10-05T03:51:17Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05750", "title": "Beyond Semantic Similarity: Performance and Costs of Agentic Retrieval for Complex Tasks", "published": "2026-10-05T03:55:10Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05774", "title": "AdaSpark: Adaptive DSpark with Online Learning for Tree Verification and N-gram Fill", "published": "2026-10-05T04:18:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05778", "title": "MedicalHarness: A Controlled Evaluation of LLMs and Agent Harnesses on Medical Tasks", "published": "2026-10-05T04:22:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05787", "title": "Constraint-Aware Conversational Job Recommendation in Code-Mixed Low-Resource Settings", "published": "2026-10-05T04:36:18Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05828", "title": "Data-Driven Personas for Survey Simulation: Insights into Simulation Alignment Across Data-Access Regimes", "published": "2026-10-05T05:30:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05833", "title": "Request Order Matters: Cache-History Sensitivity in Selective KV-Cache Reuse for Rolling Agents", "published": "2026-10-05T05:37:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05842", "title": "HLA: Expressive Hybrid Linear Attention via Chunk-Wise Dynamic Mixing", "published": "2026-10-05T05:49:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05861", "title": "Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training", "published": "2026-10-05T06:28:58Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05863", "title": "CCQ: A Multi-State Child Care Quality Dataset to Support AI for Children's Health Research", "published": "2026-10-05T06:30:09Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05872", "title": "Off-Policy Merging Beats On-Policy Self-Distillation for Continual Learning", "published": "2026-10-05T06:40:34Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.05874", "title": "Global Communication or Graph-Specific Memory?", "published": "2026-10-05T06:43:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05879", "title": "Learning to Learn a Language", "published": "2026-10-05T06:52:58Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05885", "title": "The Optimization Landscape of Learning Compacted Context Models", "published": "2026-10-05T07:00:44Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05896", "title": "TasteRoute: Personalized Routing for Video Generation", "published": "2026-10-05T07:12:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05897", "title": "Fitting Vision Adapters at Frontier Scales", "published": "2026-10-05T07:13:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05935", "title": "ThunderSyncRL: Lossless Acceleration of Agentic Reinforcement Learning", "published": "2026-10-05T07:51:46Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.05940", "title": "ReMem: Streaming Video Understanding With Long Context Retention", "published": "2026-10-05T07:56:41Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05949", "title": "Technical Report on the Turba Fertilizer Machine Learning Stack in Morocco", "published": "2026-10-05T08:04:09Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.05961", "title": "Strategic Multi-Agent Learning for Interpretable Action Valuation of All Players in Football", "published": "2026-10-05T08:13:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05978", "title": "Byte Language Models: Scaling, Emergent Abstractions, and Information Allocation", "published": "2026-10-05T08:29:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05981", "title": "Mechanizing the User's Eye: Pre-Registered Deployment of a Sabotage-Validated Fail-Plausible Observer in a Production LLM Agent Runtime", "published": "2026-10-05T08:34:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.05982", "title": "Breaking the Tie: A Cluster-Aware Routing Framework for Large Language Models", "published": "2026-10-05T08:35:05Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05993", "title": "From Transformation to Target State: Rethinking Query Representation for Zero-Shot Composed Image Retrieval", "published": "2026-10-05T08:47:15Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.05994", "title": "How (and How Not) to Use Data Augmentation in VLA Post-Training", "published": "2026-10-05T08:48:13Z", "status_at_scan": "new", "tracks": [ "post-training", "recommendation" ] }, { "arxiv_id": "2610.06001", "title": "AgentSpy: Making AI Agent Behavior Observable", "published": "2026-10-05T08:55:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06018", "title": "Investigating Query-Insensitive Behavior in Spatio-Temporal Video Grounding", "published": "2026-10-05T09:17:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06019", "title": "Adaptive Expert Guidance for Efficient On-Policy Reinforcement Learning", "published": "2026-10-05T09:18:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06021", "title": "Scalable Minimal-Change Learning for Controllable Image Editing", "published": "2026-10-05T09:18:55Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.06025", "title": "Grounded Joint-Attention Other-Play for Zero-Shot Coordination", "published": "2026-10-05T09:23:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06026", "title": "Differentiable Bit-Widths: Co-optimizing Pruning and Quantization via SVD for Ultra-Efficient LLM Compression", "published": "2026-10-05T09:25:08Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06039", "title": "MercerFlow: Flow Matching in a Kernel-Induced Latent Space for Probabilistic Forecasting", "published": "2026-10-05T09:38:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06047", "title": "Let the Agent Do It? How Software Practitioners Understand and Make Permission Decisions in Agentic AI Assistants", "published": "2026-10-05T09:43:54Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06050", "title": "MATE: Adaptive Long- and Short-Term User Memory for LLM-Based Recommendation", "published": "2026-10-05T09:44:24Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.06056", "title": "ROT: Rotating Hidden States towards Contextual Vectors for Hallucination Mitigation in LVLMs", "published": "2026-10-05T09:49:09Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06063", "title": "Vision Transformer Ensembles for Panoramic Street Segmentation", "published": "2026-10-05T09:56:48Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06064", "title": "TrustMI: Causally controlling how assistants trust their users", "published": "2026-10-05T09:56:52Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06069", "title": "Attention Tax, Handoff Tax: A Stylised Model of When Multi-Agent LLM Systems Help", "published": "2026-10-05T10:04:20Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.06083", "title": "Boosting Transferable Adversarial Attacks against Deep Reinforcement Learning", "published": "2026-10-05T10:14:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06100", "title": "From Traces to Agentic Worlds: Agentic Language World Models for Interactive Environment Simulation", "published": "2026-10-05T10:32:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06122", "title": "Benchmarking Jailbreak Guardrails for Embodied Agents", "published": "2026-10-05T10:56:18Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06147", "title": "Efficient Test-time Adaptation through Candidate Verification and Divergence Shifts", "published": "2026-10-05T11:18:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06148", "title": "Reinforcement Learning-Based Optimization of Workload-Aware Power Delivery Networks", "published": "2026-10-05T11:19:00Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06161", "title": "Introducing Code-Switched Contexts to Cognitively-Inspired Bilingual Model Training", "published": "2026-10-05T11:30:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06163", "title": "Where Did the Repair First Go Wrong? Localizing the Origins of Silent Failures in Agentic Vulnerability Repair", "published": "2026-10-05T11:31:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06190", "title": "Bridging the Evidence-to-Execution Gap:A Reflective Agent for Multi-Objective Peptide Design", "published": "2026-10-05T12:07:21Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06191", "title": "Judged Useless, Queried Anyway: Tool-Using Agents Rarely Turn Their Own Evidence Judgments into Stopping Decisions", "published": "2026-10-05T12:08:04Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06192", "title": "Copies or Sources? Measuring How LLM Aggregators Count Restated Evidence in Multi-Agent Systems", "published": "2026-10-05T12:08:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06193", "title": "Correct Code, Broken Contributions? SWE-CC: Benchmarking Repository Policy Compliance for Coding Agents", "published": "2026-10-05T12:08:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06196", "title": "EORestore-Agent: Fidelity-Guided Agentic Restoration of Remote Sensing Images with Composite Degradations", "published": "2026-10-05T12:11:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06204", "title": "Do Small Language Models Learn to Negotiate? A Controlled Scaling Study of RL-Trained Sellers", "published": "2026-10-05T12:17:22Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06207", "title": "AI-Decision Checkpoints for AI-Augmented Business Process Management: Framework and Educational Instantiation", "published": "2026-10-05T12:18:14Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06226", "title": "LeAVJEPA: A Minimalist Architecture for Audio-Visual Self-Supervised Learning", "published": "2026-10-05T12:33:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06251", "title": "Shared Stopping Decisions Change Answers in HQQ Cache Quantization", "published": "2026-10-05T12:47:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06269", "title": "Evolving in Thought Space: Training a Small Model at Test Time Unlocks Better Discoveries", "published": "2026-10-05T13:04:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06286", "title": "DeferKV: Rethinking Eviction Timing for One-Shot KV Cache Compression", "published": "2026-10-05T13:13:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06293", "title": "VepAgent: Bridging Causal-Transition via Tool-Augmented Reinforcement Learning for Video Event Prediction", "published": "2026-10-05T13:19:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06320", "title": "CRAFTER: Causality-based Self-adaptation for Autonomous IoT Systems", "published": "2026-10-05T13:34:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06324", "title": "Readout Blindness: VLM Scores Miss the Spatial Direction Their Frozen Encoders Retain", "published": "2026-10-05T13:36:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06360", "title": "Ontology Concept Overlap as a Training Signal: Knowledge-Grounded Reinforcement Learning for Clinical Question Answering", "published": "2026-10-05T13:59:39Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06387", "title": "Scaling Down the Scaling Laws: Parameter Efficiency and Compute-Optimal Training in Resource-Constrained Large Language Models", "published": "2026-10-05T14:12:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06394", "title": "Harnessing Multimodal Large Language Models for Training-Free Human-Object Interaction Detection", "published": "2026-10-05T14:16:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06401", "title": "RAISED: Self-Distillation for Robustness to Prompt Injection in LLM Agents", "published": "2026-10-05T14:21:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06404", "title": "Quantifying the Stability of Multi-Step Reasoning via Error Amplification", "published": "2026-10-05T14:22:59Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06406", "title": "What Did the Agent Actually Do? Evidence-Grounded Oversight for Long-Horizon Agents", "published": "2026-10-05T14:25:16Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.06409", "title": "FlashCart: Fast Cartesian Tensor Products for Equivariant Interatomic Potentials", "published": "2026-10-05T14:26:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06411", "title": "From Benchmark to Bench: Can Agents Survive Real-World Drug Discovery?", "published": "2026-10-05T14:28:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06413", "title": "SpatialChain: A Benchmark for Auditing Spatial Reasoning Faithfulness in VLMs", "published": "2026-10-05T14:29:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06439", "title": "Better Call Reward: Reward Hacking as Strategic Abstention in Legal Reasoning Models", "published": "2026-10-05T14:46:35Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06446", "title": "The Assistance Dilemma: Learning to Teach via Multi-Turn Reinforcement Learning", "published": "2026-10-05T14:50:58Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06454", "title": "AgentPrivArena: Evaluating and Auditing Real-world AI Agent Privacy", "published": "2026-10-05T14:55:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06479", "title": "Behavior-Preserving KV Cache Compression", "published": "2026-10-05T15:06:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06514", "title": "ANT: A Multi-Granularity Network Traffic Dataset and Benchmark for Agents Behavior Auditing", "published": "2026-10-05T15:30:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06563", "title": "HERA: Harness-Environment Co-Evolution for Reliable Agentic Abstention", "published": "2026-10-05T15:49:12Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06571", "title": "BrainTRACE: Tracing Longitudinal, Multimodal, and Volumetric Evidence in Brain MRI Clinical Reasoning", "published": "2026-10-05T15:52:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06576", "title": "Before Agent Tells The Lie: Has Deception Already Been Represented?", "published": "2026-10-05T15:54:02Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06582", "title": "Mind the Execution Gap: Action-Semantic Mismatch in World-Model Control", "published": "2026-10-05T15:59:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06584", "title": "Does AI Help Cyber Attackers or Defenders? Evidence from Nonpublic Vulnerabilities and Subsequent Attacks", "published": "2026-10-05T16:00:23Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06590", "title": "SPRIG: Semantic-ID-enhanced Paths for Knowledge Graph-based Generative Recommendation", "published": "2026-10-05T16:04:55Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.06597", "title": "Can Agent Harnesses and Inference Engines Hear Each Other? The HEAR Protocol for Agentic LLM Serving", "published": "2026-10-05T16:07:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06598", "title": "SimForcing: Distilling Simulation Motion Priors into Real-Domain Robot World Models", "published": "2026-10-05T16:08:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06603", "title": "Word-Level Text Unmixing via Evidence-Preserving Ownership Routing with Language Models", "published": "2026-10-05T16:12:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06611", "title": "Lens3D: Target-Conditioned Visual Foveation for Fine-Grained 3D Understanding", "published": "2026-10-05T16:16:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06614", "title": "FREA: A Multi-Source Expert Benchmark for Reaction Feasibility Verification", "published": "2026-10-05T16:17:37Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06616", "title": "Video Encoders Built on Image Representations", "published": "2026-10-05T16:18:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06617", "title": "RealtimeWAM: One-Step Asynchronous World Action Models", "published": "2026-10-05T16:18:07Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06636", "title": "Differentially Private Mixing of Public Datasets Improves Private Learning", "published": "2026-10-05T16:25:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06637", "title": "Long-Horizon Textual World Modeling through Structured Reasoning", "published": "2026-10-05T16:25:44Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06647", "title": "LoGRA: Scaling LLM Reinforcement Learning with Low-Rank Gradient Sketches", "published": "2026-10-05T16:28:11Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06648", "title": "Representation-Space MMD for Diffusion Language Models", "published": "2026-10-05T16:28:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06650", "title": "Wikidata Search Traces: A Dataset for Training Knowledge Graph Search Agents", "published": "2026-10-05T16:29:22Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.06651", "title": "Considering Context: When World Models Need Context Encoders", "published": "2026-10-05T16:32:15Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06666", "title": "What Matters for Latent Reasoning with Flow Matching", "published": "2026-10-05T16:42:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06670", "title": "Reward Stealing Attack on Large Language Models", "published": "2026-10-05T16:44:00Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06671", "title": "Learning What to Imitate: Entropy-Aware Distribution Mixing", "published": "2026-10-05T16:44:02Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06685", "title": "Aligning Multimodal Patient Evidence with Biomedical Knowledge Graphs for Clinical LLMs", "published": "2026-10-05T16:51:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06686", "title": "OVAL: Output-Aware Local Page Bases for KV Cache Retrieval", "published": "2026-10-05T16:51:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06695", "title": "MedPrune: Topology-Efficient Multimodal Multi-Agent Communication Evolution for Medical VQA Tasks", "published": "2026-10-05T16:55:58Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06703", "title": "Reading the Mood: Emotion-Guided Book-to-Music Recommendation via CGANs and LLMs", "published": "2026-10-05T17:01:18Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.06715", "title": "Domain adaptation of Russian ModernBERT for long legal documents", "published": "2026-10-05T17:07:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06726", "title": "Decoupling Time and Space: A Temporally Conditioned Refinement for EEG Source Imaging", "published": "2026-10-05T17:15:47Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06729", "title": "Improving Diversity in LLM Short Story Generation", "published": "2026-10-05T17:16:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06744", "title": "ufakzeka-karar: An Open Turkish Typed-Decision Model with Order-Invariant Option Scoring", "published": "2026-10-05T17:22:23Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06748", "title": "BazaarBench: Delegation Safety in Decentralized C2C Marketplaces Run by LLM Agents", "published": "2026-10-05T17:26:17Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06750", "title": "Balancing Memory Pathways: Analyzing and Improving Memory Utilization in Hybrid LMs", "published": "2026-10-05T17:27:30Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.06751", "title": "MatrixFormer: A Foundation Model for Matrix Completion", "published": "2026-10-05T17:28:38Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.06782", "title": "T-Search: An Open Agentic Retriever and Playground for Hard Multi-Step Search", "published": "2026-10-05T17:44:26Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06784", "title": "How to scale your HEP ML models: A recipe for robust architecture comparisons at scale", "published": "2026-10-05T17:44:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06790", "title": "Back to the Future: Rethinking EDA Infrastructure for Agentic Systems in Chip Design Verification", "published": "2026-10-05T17:48:14Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06804", "title": "Sharpen Without Search: On-Policy Distillation of Sequence-Level Power Distribution", "published": "2026-10-05T17:53:30Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06817", "title": "Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model", "published": "2026-10-05T17:56:07Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06825", "title": "PlotGround: Grounding Plot Digitization in Real Scientific Figures and Their Source Data", "published": "2026-10-05T17:57:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06829", "title": "CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling", "published": "2026-10-05T17:57:51Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.06830", "title": "MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents", "published": "2026-10-05T17:58:06Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06833", "title": "Towards Looped Models Done Right, Part II: Rethinking at Fixed Points", "published": "2026-10-05T17:58:15Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06843", "title": "Recursive Video In-Context Learning for Agentic Robot", "published": "2026-10-05T17:59:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06844", "title": "Learning to Read the Contextual Tokens in Diffusion Transformers", "published": "2026-10-05T17:59:17Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06851", "title": "Base Models Can Reason By Taking a Cue From Training Data", "published": "2026-10-05T17:59:52Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06874", "title": "Statistical Turbulence and High-Fidelity Disturbance Fields for Quadrotor Flight Control", "published": "2026-09-09T05:14:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06892", "title": "Axiom Satisfiability of Linear Rewards in Alignment", "published": "2026-09-26T15:24:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06896", "title": "Medical Image Alignment Assessment as a Test of Generalist Visual Reasoning in Frontier Multimodal Models", "published": "2026-09-28T17:17:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06902", "title": "Tree Navigation Without LLM Summaries: A Matched-Cost Study of Hierarchical Retrieval for Long-Document QA", "published": "2026-09-30T08:42:45Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06910", "title": "GAMEGO: Training Game-Dev Agents with Synthetic Trajectories Anchored in Real-World Assets", "published": "2026-10-02T17:05:07Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06918", "title": "Learning from Unreliable Trajectories: Adversarially-Robust Federated Q-Learning", "published": "2026-10-02T20:54:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06923", "title": "RadOnc-Agent: An LLM-Orchestrated Framework for AI Workflows Across the Radiotherapy Care Pathway", "published": "2026-10-02T22:39:48Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06932", "title": "RADC: Risk-Aware Dual Caching for Vision-Language Test-Time Adaptation", "published": "2026-10-03T02:28:11Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06947", "title": "FactorBench: A Portfolio-Aware Benchmark for Automated Factor Mining", "published": "2026-10-03T10:57:37Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.06949", "title": "AdaLoop: Adaptive-Depth Latent Reasoning for Audio Language Models", "published": "2026-10-03T11:27:41Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06950", "title": "Learning to Decide, Not to Reason: Parameter-Efficient Decision Operators via Low-Rank Activation Steering", "published": "2026-10-03T12:57:57Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.06962", "title": "Verdicts Without Annotated Evidence: Rejection Sampling or Label-Only Post-Training for Evidence Recovery?", "published": "2026-10-03T17:01:42Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.06963", "title": "WavePrune: One period is often enough for RoPE", "published": "2026-10-03T17:01:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06964", "title": "Principles that Guide, Actions that Inform: Agent Evolution via Knowledge Abstraction", "published": "2026-10-03T17:29:38Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06966", "title": "APEX: Active Protection at Execution Boundaries for LLM Agents", "published": "2026-10-03T17:53:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06971", "title": "AegisFlow: A Multi-Agent Agentic AI Framework for Autonomous Remediation and Self-Healing in Fragile Data Ecosystems", "published": "2026-10-03T19:12:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.06972", "title": "BoT-Feedback: Grounding Multimodal Reasoning in Biomechanical Evidence for Explainable Human Action Feedback", "published": "2026-10-03T20:22:29Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06977", "title": "Visual-Invariance-Augmented Feature Optimal Alignment for Transferable Adversarial Attacks against Closed-Source MLLMs", "published": "2026-10-04T00:16:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.06993", "title": "DART-ES: Difficulty-Aware Reweighting and Targeted Replay for Fine-Tuning LLMs with Evolution Strategies", "published": "2026-10-04T07:42:11Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07004", "title": "Topology-Consistent Task Planning over Cellular Workflow Complexes for LLM-based Agents", "published": "2026-10-04T10:24:20Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07014", "title": "DTFormer: Text-Guided Semantic Alignment for RGB-D Segmentation", "published": "2026-10-04T14:06:32Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07023", "title": "Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment", "published": "2026-10-04T15:50:42Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07032", "title": "Investigating Model Compression for Neural Machine Translation in the Biomedical Domain", "published": "2026-10-04T19:01:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07043", "title": "TRIAGE: Direction-Aware Mismatch Stabilization of Native NVFP4 Reinforcement Learning", "published": "2026-10-05T00:53:11Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07060", "title": "Skillful Data-Driven Subseasonal Soil Moisture Forecasting: Prospects and Limits for Flash Drought Prediction", "published": "2026-10-05T06:34:10Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07062", "title": "Learning to Simulate Individuals from Macro Social Signals", "published": "2026-10-05T06:56:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07086", "title": "SchemaFill: Efficient LLM Tool Calling via Slot-Parallel Speculative Decoding", "published": "2026-10-05T12:22:34Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07089", "title": "Towards a Unified Misuse Monitoring Benchmark", "published": "2026-10-05T12:57:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07091", "title": "Smart Content Ingestion for Generative AI Workloads", "published": "2026-10-05T13:22:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07105", "title": "Beyond Successor Accuracy: State Retention for Recursive Self-Improvement in Recommendation", "published": "2026-10-05T15:03:27Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.07118", "title": "AMBER: Training Long-Horizon Web Agents through Append-Only Memory", "published": "2026-10-05T17:11:07Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07121", "title": "SoloQ: Calibration-Free Quantization for Diffusion Language Models", "published": "2026-10-05T17:30:54Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07127", "title": "PlaySuite: A Large-Scale Benchmark for Interactive Visual Intelligence", "published": "2026-10-05T17:52:40Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07132", "title": "CroissantMiner: Automated Extraction and Validation of Croissant Metadata for ML Datasets", "published": "2026-10-05T17:59:48Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07168", "title": "A theory of platonic representations in language models", "published": "2026-10-05T18:00:09Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07177", "title": "CLM-as-a-Judge: Evaluating an Open Contrastive Decision Model on Public Judge Benchmarks", "published": "2026-10-05T18:01:37Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07191", "title": "Agentic Design Space Exploration for Joint Hardware Configuration Selection and Mapping of AI Inference Workloads on Heterogeneous Edge SoCs", "published": "2026-10-05T18:08:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07212", "title": "Reward-Driven Learning under Prompt-Level Differential Privacy", "published": "2026-10-05T18:26:21Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07226", "title": "Minimal Witness Reinforcement Learning", "published": "2026-10-05T18:34:49Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07243", "title": "Hybrid Cross-Modal Attention Network for Early Breast Cancer Detection in Low-Resource Clinical Settings", "published": "2026-10-05T18:44:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07247", "title": "Learning What to Distill: Bilevel Top-K Token Selection for Self-Distillation in Large Language Models", "published": "2026-10-05T18:48:00Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07258", "title": "Lineage-Aware Memory Governance: A Derivation-Gated Framework for Privacy-Preserving Column-Level Access Control in Enterprise AI Agents", "published": "2026-10-05T18:56:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07261", "title": "Verifying Coordination in Parallel Coding Agents: NP-Bench and a Scheduling Planner", "published": "2026-10-05T19:02:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07269", "title": "What Words Keep of a Place: Zero-Shot Language Reasoning for Cross-View Geo-Localization", "published": "2026-10-05T19:11:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07289", "title": "Catching Developers in the Flow: Low-Latency Agentic Program Repair at Google Scale", "published": "2026-10-05T19:23:06Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07311", "title": "Understanding and Mitigating Inference-Time Overreliance Using Agentic Memory", "published": "2026-10-05T19:48:54Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07324", "title": "Scale-Invariant Training for Time Series Foundation Models", "published": "2026-10-05T19:56:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07332", "title": "Structuring MoE Expert Selection for Agentic Reinforcement Learning", "published": "2026-10-05T20:06:47Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07335", "title": "Selective Critique for Cost-Aware LLM Agents in Long-Horizon Decision Making", "published": "2026-10-05T20:09:48Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07338", "title": "Logbook: Extremely Long-form Audio Event Understanding", "published": "2026-10-05T20:13:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07339", "title": "A doctrine-grounded visual question answering dataset for Tactical Combat Casualty Care", "published": "2026-10-05T20:14:31Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07348", "title": "Stepped MoE: Segment-Level Routing with Configurable Inference Complexity", "published": "2026-10-05T20:18:26Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07349", "title": "RELACE: retrospective likelihood-based action credit estimation for long-horizon language agents", "published": "2026-10-05T20:19:04Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07355", "title": "Tracking Is Not Permanence: What Video World Models Keep of a Hidden Object", "published": "2026-10-05T20:22:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07359", "title": "Evaluate the Stack, Not the Layer: Do Deterministic and LLM Gates for Agent Actions Fail Independently?", "published": "2026-10-05T20:26:58Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07376", "title": "MemCo: Memory-Centric Collaboration for Generalizing LLM Agents to Unseen Environments", "published": "2026-10-05T20:46:22Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07385", "title": "Semantic Capability Acquisition and Specialization During Vision-Language Model Fine-Tuning", "published": "2026-10-05T20:56:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07389", "title": "Inference and learning in sparse autoencoders as natural gradient flow", "published": "2026-10-05T21:03:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07402", "title": "Rethinking Semantic ID Construction for Generative Recommendation: SimHash with Parallel Decoding and Semantic Alignment", "published": "2026-10-05T21:15:13Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.07423", "title": "2d-fet-bench: from spatial reasoning to fet design on flakes", "published": "2026-10-05T21:30:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07426", "title": "AccentCL: Robust Accent Classification with Incremental Expansion", "published": "2026-10-05T21:34:23Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07457", "title": "AlignQuant: Tile-Aligned Mixed-Precision Quantization for Efficient LLM Generation", "published": "2026-10-05T22:06:04Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07460", "title": "ElasticFit: Fit-Aware 3D Object Insertion via VLM Reasoning and Generative Adaptation", "published": "2026-10-05T22:12:18Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07473", "title": "PsyCIDRA: A Dual-Agent Framework for Psychiatric Interviewing and Diagnostic Reasoning", "published": "2026-10-05T22:40:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07475", "title": "Adapting to Changes in Agent Behavior via Finite-Depth Policy Sensitivity", "published": "2026-10-05T22:40:40Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07489", "title": "Deep Defence on Wheels: A Dual Intrusion Detection System Architecture for Comprehensive In-Vehicle Network Security", "published": "2026-10-05T22:52:39Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07491", "title": "Who Bears the Burden? Learning Responsibility for Shared Constraints in Multi-Agent Reinforcement Learning", "published": "2026-10-05T22:53:33Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07509", "title": "On Open-Ended Information Seeking for Information Elicitation Agents", "published": "2026-10-05T23:20:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07518", "title": "Harmful SFT Leaves a Continuous Trace in LLM Checkpoint Updates", "published": "2026-10-05T23:28:45Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07522", "title": "Activation Denoising: A Robustness View on Parallel vs Sequential LLM Quantization", "published": "2026-10-05T23:41:48Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07532", "title": "Safeguarding LLMs via Model-Agnostic Latent Safety Signals from Dark Knowledge", "published": "2026-10-05T23:59:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07533", "title": "SkillFormer: Skill-Decomposed Adaptation for Audio Language Models", "published": "2026-10-06T00:00:21Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07534", "title": "The EPIC Framework for Spec-Driven Development", "published": "2026-10-06T00:00:54Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07535", "title": "Disentangling Models from Personas in Heterogeneous LLM Simulations", "published": "2026-10-06T00:01:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07545", "title": "Quality-Aware Self-Correcting Speech Translation on an Edge Device", "published": "2026-10-06T00:21:08Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07550", "title": "Foundation Model-Aided Multi-Agent Reinforcement Learning for Wireless Random Access Network Optimization", "published": "2026-10-06T00:27:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07553", "title": "Which and When to Admit: Gradient Admission for Data-Centric Small Language Model Finetuning", "published": "2026-10-06T00:34:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07556", "title": "Decoupled Multi-Agent Orchestration", "published": "2026-10-06T00:41:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07557", "title": "CheckerBench: Can Long-Horizon Agents Synthesize Static-Analysis Checkers?", "published": "2026-10-06T00:42:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07569", "title": "OpenSplatGraph: From Dense Semantic Maps to Structured Scene Graphs for Open-Vocabulary Robot Perception", "published": "2026-10-06T00:58:30Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07570", "title": "Unanimously Wrong: Certified Abstention from How Medical LLM Consensus Forms", "published": "2026-10-06T00:59:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07572", "title": "Two Vectors Replace In-Context Demos: Structured Task Adaptation via Embeddings", "published": "2026-10-06T01:05:57Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07576", "title": "CETUS: How Far Do Representations Trained on Earth Transfer to Cassini SAR of Titan?", "published": "2026-10-06T01:11:34Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07578", "title": "Cooperating with Future Collaborators: Multi-Agent RL under Staggered Participation", "published": "2026-10-06T01:13:24Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07585", "title": "REViT-v2: Hierarchical Windowed Roto-reflection Equivariant ViT for Equivariant Feature Extraction", "published": "2026-10-06T01:23:15Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07587", "title": "Large Language Model Orchestration under Heterogeneous Preferences via Explicit Persona Inference", "published": "2026-10-06T01:24:56Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.07588", "title": "Personal-Agent Mediated Recommendation with Cross-Platform User History", "published": "2026-10-06T01:26:01Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07594", "title": "BiGym 2.0: Benchmarking Learned and Agent-Developed Policies for Humanoid Household Manipulation", "published": "2026-10-06T01:33:35Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07615", "title": "AFA-BANDIT: Provably Near-Optimal Online Multi-Feature Classification Under Budget Constraints", "published": "2026-10-06T02:02:27Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07625", "title": "Stateless Language Agents: Scaling Long-Horizon Automated Research", "published": "2026-10-06T02:17:32Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07639", "title": "HarnessSecurity-Bench: Do Security Mechanisms Really Protect Coding Agent Harnesses?", "published": "2026-10-06T02:34:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07640", "title": "Towards the Automatic Synthesis of Interpretable Chess Tactics", "published": "2026-10-06T02:35:09Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07643", "title": "Monte Carlo Estimation for KV Cache Eviction", "published": "2026-10-06T02:36:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07645", "title": "SkillPoison: Progressive Skill Poisoning via Successful Experiences", "published": "2026-10-06T02:38:36Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07654", "title": "Does On-Policy Distillation for Safety Pose Backdoor Risks?", "published": "2026-10-06T02:50:07Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07657", "title": "Where Rules End and Judges Begin: Measuring the Judgment Boundary in Multi-Agent Systems Security", "published": "2026-10-06T02:52:51Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07668", "title": "CACHEFORGE: LLM-Guided End-to-End Generative Cache Replacement Policy for Performance and Hardware Efficiency", "published": "2026-10-06T03:03:13Z", "status_at_scan": "new", "tracks": [ "foundation-model", "post-training" ] }, { "arxiv_id": "2610.07689", "title": "Unlocking Fine-Grained Perception in CLIP via Structurally-Aware Latent Masked Modeling", "published": "2026-10-06T03:25:22Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07704", "title": "Independent Multi-Agent Reinforcement Learning with Counterfactual Semantic-Social World Models", "published": "2026-10-06T03:53:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07706", "title": "WASD: Wasserstein-based Knowledge Distillation for Large Language Models", "published": "2026-10-06T03:58:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07729", "title": "Foveated Compression: Selective High-Resolution Preservation for Token-Efficient VLMs", "published": "2026-10-06T04:28:13Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07730", "title": "SanSi: A Looped Typed Decision Model for System 1.5 Thinking", "published": "2026-10-06T04:28:22Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07731", "title": "Learning to Retrieve via Reinforcement Learning in Embedding Space", "published": "2026-10-06T04:29:29Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07739", "title": "Cite What You Explore: Budget-Aware LLM Reasoning over Medical KGs with Verifiable Evidence", "published": "2026-10-06T04:36:07Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07751", "title": "How Well Do LLMs Reason with Noisy Evidence? An Active Visual Reasoning Benchmark", "published": "2026-10-06T04:49:11Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07753", "title": "From Evidence to Action: How Tool-Using Agents Fail", "published": "2026-10-06T04:50:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07755", "title": "Trustworthy Method Comparison with AI Judges: Estimation and Design under Order, Batch, and Aggregation Effects", "published": "2026-10-06T04:53:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07757", "title": "Acquiring and Verifying Repository Norms for Coding Agents", "published": "2026-10-06T04:53:48Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07758", "title": "Later Is Better: Token Reduction for ViTs Under Distribution Shift", "published": "2026-10-06T04:54:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07761", "title": "Contrastive Learning for Aspect Representation towards Explainable Recommendation", "published": "2026-10-06T04:55:55Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.07763", "title": "ST-Bench: A Spatial-Temporal Benchmark for Multi-Agent System Generation on Scientific Research Tasks", "published": "2026-10-06T04:59:49Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07767", "title": "TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models", "published": "2026-10-06T05:03:47Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07774", "title": "Reading, Not Manipulating: Leveraging Router Logits for Multimodal Safety in MoE Vision-Language Models", "published": "2026-10-06T05:09:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07781", "title": "Quantization Effects on Tool-Failure Recovery Vary Across Prompts and Evaluation Designs", "published": "2026-10-06T05:17:10Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.07782", "title": "Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell", "published": "2026-10-06T05:17:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07787", "title": "OOPMAS: Object-Oriented Multi-Agent Systems for Query-Level Workflow Generation", "published": "2026-10-06T05:31:29Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07809", "title": "MASKerade: Token-Routed Mask Experts for Dense-to-MoE Upcycling", "published": "2026-10-06T06:04:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07810", "title": "SIFT: Search Intent-to-Filter Transformer for Multi-Task Personalized Filter Ranking at Airbnb", "published": "2026-10-06T06:05:45Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.07816", "title": "Do I Need the Cloud? Uncertainty-Aware Step-Level Handoff for Small Language Model Agents", "published": "2026-10-06T06:09:31Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07817", "title": "One Step at a Time: Trading LLM Autonomy for Process Predictability", "published": "2026-10-06T06:10:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07819", "title": "$α$Transfer: Coefficient Transfer for Efficient Model Merging", "published": "2026-10-06T06:13:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07825", "title": "Forecast Accuracy Is Not Trading Profit: Evolving Small Recurrent Networks for Stock Return Prediction", "published": "2026-10-06T06:20:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07832", "title": "Harness Engineering for Software Engineering via Modular Executable Dev-Primitives", "published": "2026-10-06T06:32:11Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07835", "title": "DHCG: Dynamic Construction of Hierarchical Collaboration Graphs for LLM-Based Multi-Agent Reasoning", "published": "2026-10-06T06:33:40Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07842", "title": "Privileged Context as Drift in On-Policy Self-Distillation", "published": "2026-10-06T06:45:39Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.07848", "title": "Dynamic Positional Attention Modulation for Parameter-Efficient Fine-Tuning of Large Language Models", "published": "2026-10-06T06:55:38Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07851", "title": "RA-MoWE: Workflow-Affinity Embeddings for Query Clustering and Agentic Workflow Generation", "published": "2026-10-06T06:58:16Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07860", "title": "WorkflowOps: Learning Agent Collaboration Priors for Multi-Agent Workflow Orchestration", "published": "2026-10-06T07:01:38Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07862", "title": "A self-learning scientific agent for X-ray diffraction", "published": "2026-10-06T07:09:39Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07863", "title": "ReFold: Training-Free Reversible Inter-Turn Context Folding for Long-Horizon Agents", "published": "2026-10-06T07:09:54Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07868", "title": "CueRator: Agentic Search for Symbolic Rules to Adapt Frozen Multimodal Encoders", "published": "2026-10-06T07:18:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07881", "title": "Self-Referenced Social Preferences: Cooperation without Observing Others Rewards", "published": "2026-10-06T07:29:38Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07886", "title": "ShanLiangRen: A Nutrition Agent for Personalized Daily Meal Planning", "published": "2026-10-06T07:32:50Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07887", "title": "Visual Abstention in Unified Multimodal Models", "published": "2026-10-06T07:33:42Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07898", "title": "FC-SWE: Failure-Conditioned RL for Long-Horizon Software Engineering Agents", "published": "2026-10-06T07:44:12Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.07903", "title": "Unsupervised Long-Tailed Adaptation of Vision-Language Models", "published": "2026-10-06T07:49:49Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07913", "title": "Multimodal Knowledge Distillation for Gastric Adenocarcinoma Classification from Whole-Slide Images", "published": "2026-10-06T07:56:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07925", "title": "TF-PRVR: Training-Free Partially Relevant Video Retrieval", "published": "2026-10-06T08:03:58Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07935", "title": "SIGMA: Self-Improving Alignment Generalization from a Model Spec", "published": "2026-10-06T08:09:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07937", "title": "Leveraging a four-quadrant approach for evaluating Redpine Science", "published": "2026-10-06T08:11:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07940", "title": "Hybrid Latent Attention for Looped Language Models", "published": "2026-10-06T08:14:07Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07948", "title": "Confidence Reasoning Graphs: Structured Confidence Estimation for LLM Agents", "published": "2026-10-06T08:22:14Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.07954", "title": "Revisiting Numerical Forecasting Models for Language-Based Trajectory Prediction", "published": "2026-10-06T08:28:52Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07966", "title": "Feature Encoding in VAE-based Audio Decoders: Effects of Input, Depth and Distribution", "published": "2026-10-06T08:32:02Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07969", "title": "EmbodiedSmith: Scaling Embodied Data through Recursive Self-Improvement Flywheel in Simulation", "published": "2026-10-06T08:34:28Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07972", "title": "Can Agents Work for Everyone? Cross-User Reliability for Mobile GUI Agents in Personalized User Interfaces", "published": "2026-10-06T08:38:30Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.07982", "title": "M3SunAgent: Monocular 3D Spatial Understanding Agent for Metric Depth Estimation and 3D Visual Grounding", "published": "2026-10-06T08:45:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07987", "title": "VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs", "published": "2026-10-06T08:48:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.07990", "title": "A Broader Look at Model Merging: Rethinking Implicit Regularization Induced by Task Arithmetic", "published": "2026-10-06T08:49:40Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08018", "title": "Structured but Silent: Probing Capability Requirements in LLM Hidden States", "published": "2026-10-06T09:12:53Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08048", "title": "DAEDALUS: Bootstrapping Agent Memory from Self-Generated Tasks", "published": "2026-10-06T09:46:32Z", "status_at_scan": "new", "tracks": [ "agent", "foundation-model" ] }, { "arxiv_id": "2610.08068", "title": "PhysTacGen: Physics-Aware Visual-Tactile Sensor Image Generation", "published": "2026-10-06T10:00:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08076", "title": "SpeedrunBench: Challenging LLM Agents with Video Game Speedrunning", "published": "2026-10-06T10:06:25Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08077", "title": "Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight", "published": "2026-10-06T10:06:47Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.08082", "title": "POLAR: Ontology-Guided Risk Prevention for Tool-Calling LLM Agents", "published": "2026-10-06T10:14:16Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.08090", "title": "Explainable Rule Mining of IPv6 Extension-Header Presence Patterns from Paired-Vantage Captures", "published": "2026-10-06T10:21:12Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08097", "title": "When Tools Lie: Reliability of Mathematical Agents Under Corrupted Tool Feedback", "published": "2026-10-06T10:28:19Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08101", "title": "Beyond Corrected Memory: Execution Consistency in Multi-Agent Systems", "published": "2026-10-06T10:29:41Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08106", "title": "ChartBmkAgent: Harness-Governed Multi-Agent Construction of Chart QA Benchmarks from Sparse Error-Taxonomy Specifications", "published": "2026-10-06T10:31:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08108", "title": "Enhancing Diffusion Language Models with Autoregressive Post-Training Weights", "published": "2026-10-06T10:31:54Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.08129", "title": "Do LLMs Act on What They Know? From Partner Representations to Cooperative Actions", "published": "2026-10-06T10:44:55Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08133", "title": "VLA-ACL: Action-Consistent Visual Token Pruning for Efficient Vision-Language-Action Models", "published": "2026-10-06T10:47:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08136", "title": "Adapting Generative Recommenders for Multi-Turn Interaction", "published": "2026-10-06T10:51:53Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08155", "title": "Token-Efficient Multi-Agent Collaboration via System One-Guided Computational Division of Labor", "published": "2026-10-06T11:11:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08159", "title": "Making COMET Comparable Across Scripts: Diagnosis and Correction of Tokeniser-Induced Script Bias in Indic MT Evaluation", "published": "2026-10-06T11:11:43Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08162", "title": "The Failure Is in the Readout: Fine-Grained Emotion Recognition Benchmarks Measure Elicitation, Not Perception", "published": "2026-10-06T11:14:23Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08164", "title": "Align, Then Correct: Training-Free Two-Stage Low-Rank Compensation for Extremely Quantized Large Language Models", "published": "2026-10-06T11:16:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08183", "title": "Compact Robot Policies Need Fine-Grained Visual Representations", "published": "2026-10-06T11:36:19Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08200", "title": "Finding the Heads and the Neurons Responsible for Network Information Retrieval in Language Models", "published": "2026-10-06T11:51:03Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08208", "title": "STRUCTURALCOST: A controlled reading time dataset for modeling human sentence processing difficulty", "published": "2026-10-06T12:00:36Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08215", "title": "Learn2Play Bench: How Well Do LLM Agents Learn from Experience in Unfamiliar Environments?", "published": "2026-10-06T12:08:45Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08232", "title": "Behavior-Mining, Generative Conversations, and Collaborative Advisory: the Future of Travel and Tourism Recommender Systems", "published": "2026-10-06T12:18:15Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08245", "title": "Aligning Performance with Contribution: Towards Contribution-Aware Fair Recommendation", "published": "2026-10-06T12:26:50Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08246", "title": "LeanPlan: Optimal Planning with LLM-Generated Heuristics and Admissibility Proofs", "published": "2026-10-06T12:27:34Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08250", "title": "MASC: A Multi-Agent Self-Calibration Framework with Latent Construct Alignment for Consistent Client Role-Playing in Psychological Counseling", "published": "2026-10-06T12:31:05Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08300", "title": "Memory Depth and Reconstructed Context Width: A Controlled Evaluation of Hierarchical Retrieval", "published": "2026-10-06T13:04:43Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08327", "title": "MedZERO: Self-Evolving Agents for Open-Ended Medical Reasoning Through Controlled Knowledge Accumulation", "published": "2026-10-06T13:25:55Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08331", "title": "Transferable Spatial Temporal Coherence Adversarial Attack on Black-Box Vision Language Models for Autonomous Driving", "published": "2026-10-06T13:30:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08337", "title": "Machine Learning for German Redispatch Forecasting under Data Delays and Temporal Distribution Shift", "published": "2026-10-06T13:35:50Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08341", "title": "DIPrune: Task-Aware Token Pruning with Dual Importance for Efficient Multimodal Language Models", "published": "2026-10-06T13:38:27Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08364", "title": "Transect: Retaining Observability for Long-Horizon LLM Agent Evaluations", "published": "2026-10-06T13:52:15Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08400", "title": "Atom-JEPA: Joint-Embedding Predictive Architecture for 3D Atomistic Systems", "published": "2026-10-06T14:13:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08401", "title": "GeoPID: Decomposing and Steering Visual Information in Vision-Language Models", "published": "2026-10-06T14:14:25Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08403", "title": "SSR: Sparse Segment Reduction for Ternary GEMM Acceleration", "published": "2026-10-06T14:15:12Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08407", "title": "Seeing the Context: Enhancing Recommender Systems with Image-Derived Contextual Signals", "published": "2026-10-06T14:15:48Z", "status_at_scan": "new", "tracks": [ "recommendation" ] }, { "arxiv_id": "2610.08414", "title": "Image Bitstream Fine-grained Understanding for Privacy-Friendly AIoT", "published": "2026-10-06T14:18:54Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08430", "title": "NeMo-DCR: Bit-Exact Delta-Compressed Refit for Scalable Agentic RL at Trillion-Parameter Scale", "published": "2026-10-06T14:30:06Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08452", "title": "Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents", "published": "2026-10-06T14:38:15Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08463", "title": "UNREAL: Unifying Retrieval and Long-Context with a Single Model", "published": "2026-10-06T14:48:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08482", "title": "Knee3DVLM: Dual-Sequence Full-Volume Vision-Language Modeling for Comprehensive Knee MRI Assessment", "published": "2026-10-06T14:58:53Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08513", "title": "Wiki-Talkie: Multilingual Benchmarking of Persona-Based Agents on Real-World Discussions", "published": "2026-10-06T15:17:46Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08514", "title": "How Much Evidence Should a Coding Agent's Self-Correction Carry? Adaptive Dirichlet Evidence for Self-Distillation", "published": "2026-10-06T15:17:56Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08526", "title": "WareFly-VLA: A Vision-Language-Action Framework for UAV Navigation and Human Tracking in Smart Warehouses", "published": "2026-10-06T15:21:14Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08528", "title": "MedCORE: Criteria-Grounded Clinical Reasoning for Interpretable Medical Image Diagnosis", "published": "2026-10-06T15:23:01Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08539", "title": "RSJEV: Discriminative Remote Sensing Scene Classification with Multimodal Large Language Models", "published": "2026-10-06T15:28:36Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08560", "title": "Have I Seen Enough? Frozen Video-Language Models Encode Evidence Readiness", "published": "2026-10-06T15:43:04Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08561", "title": "Reinforcement Learning for Hierarchical Reasoning Rewards: Minimax-Optimal Rates with Transformers", "published": "2026-10-06T15:43:08Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.08586", "title": "MINDSET: Energy-based Schema Evolution for Long Conversational Agent Memory", "published": "2026-10-06T15:55:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08621", "title": "Recursive Game Creator: An Agentic Product-Level Experience-Oriented Game Harness", "published": "2026-10-06T16:21:44Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08622", "title": "Agentic RCA for Internet-Scale Services Using Constrained Creativity", "published": "2026-10-06T16:21:57Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08630", "title": "Towards In-Parameter Memory Augmentation for Large Language Models", "published": "2026-10-06T16:27:47Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08639", "title": "Forensic Reserve: Eliciting Latent Knowledge for Image Forgery Detection", "published": "2026-10-06T16:33:00Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08647", "title": "SquidAgent: Parallelize Wisely, Coordinate Efficiently", "published": "2026-10-06T16:35:28Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08649", "title": "Stable Scores, Unstable Answers: Frame Phase and Option Order in Video Multiple-Choice Evaluation", "published": "2026-10-06T16:36:06Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08651", "title": "A Case Study in Assuring AI-Written Software", "published": "2026-10-06T16:37:08Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08659", "title": "Selective Transfer of RL Updates for Visual Reasoning", "published": "2026-10-06T16:43:55Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08662", "title": "ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding", "published": "2026-10-06T16:46:14Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08668", "title": "Semantic Behavioral Watermarking: Paraphrase-Robust and Forgery-Resistant Provenance for LLM Agents", "published": "2026-10-06T16:50:42Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08670", "title": "Principled Under Pressure: Post-Training Decides Whether LLMs Act on Their Own Moral Judgment", "published": "2026-10-06T16:52:23Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.08674", "title": "EC-RAG: Event Chain Retrieval-Augmented Generation for Long Video Understanding", "published": "2026-10-06T16:57:46Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08678", "title": "Secure Speculative Decoding for Large Language Models", "published": "2026-10-06T16:59:37Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08691", "title": "ScienceClaw: Benchmarking Continual Self-Evolution of AI-for-Science Agents Across the Natural and Social Sciences", "published": "2026-10-06T17:08:03Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08713", "title": "SpaTime: Streaming Vision-Language Models for Spatio-temporal Reasoning", "published": "2026-10-06T17:22:33Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08718", "title": "When Forgetting is not Catastrophic: On the Mechanics of Spurious Forgetting", "published": "2026-10-06T17:25:48Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08720", "title": "WorldSolver: Can LLM Agents Simulate the Physical Dynamics via Solver Generation?", "published": "2026-10-06T17:26:26Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08773", "title": "AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model", "published": "2026-10-06T17:56:43Z", "status_at_scan": "new", "tracks": [ "agent", "post-training" ] }, { "arxiv_id": "2610.08775", "title": "Agent in a Bottle: Can LLM Agents Turn Their Capabilities Into Cheap, Scalable Artifacts?", "published": "2026-10-06T17:57:19Z", "status_at_scan": "new", "tracks": [ "agent" ] }, { "arxiv_id": "2610.08778", "title": "Sherpa: Teaching LLMs to Teach Adaptively", "published": "2026-10-06T17:58:18Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.08779", "title": "ALIVE: Interaction-Aligned Object Insertion for First-Frame-Guided Video Editing", "published": "2026-10-06T17:58:51Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] }, { "arxiv_id": "2610.08781", "title": "IdeaAnchor: Teaching LLMs to Turn Literature into Research Ideas", "published": "2026-10-06T17:59:01Z", "status_at_scan": "new", "tracks": [ "post-training" ] }, { "arxiv_id": "2610.08791", "title": "World Models' Last Exam in Physics", "published": "2026-10-06T17:59:56Z", "status_at_scan": "new", "tracks": [ "foundation-model" ] } ] }