{ "scope": "Read-only local reconciliation; no implementation, training, source discovery, or claim of full capability reproduction.", "root": ".", "counts": { "ledger_batches": 55, "ledger_candidate_rows": 4943, "ledger_unique_identities": 4821, "ledger_repeated_identities": 113, "manifest_rows": 930, "manifest_arxiv_identities": 921, "old_status_implementation_pending": 521, "accepted_register_union": 534, "accepted_catalogued": 26, "accepted_non_catalogued_terminal": 4, "accepted_non_catalogued_no_terminal": 504, "accepted_unclosed_missed_by_old_pending": 14, "legacy_deferred_promoted_to_pending": 31, "pending_duplicate_ledger_ids": 0, "pending_manifest_matches": 0 }, "accepted_origins": { "oct07+oct12": 494, "oct12": 26, "oct15": 14 }, "accepted_unclosed_by_track": { "foundation-model": 212, "post-training": 85, "agent": 206, "recommendation": 1 }, "accepted_unclosed": [ { "id": "2510.15831", "title": "VISTA: A Test-Time Self-Improving Video Generation Agent", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2602.16490", "title": "From Growing to Looping: A Unified View of Iterative Computation in LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00010", "title": "Heavy-Tailed Memory Traces in Long-Horizon Language Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00012", "title": "When Do Causal World Models Help Modular LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00025", "title": "Measuring the Microtask Eligibility Gap: When Is an Off-the-Shelf SLM Enough for an Agent Harness?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00040", "title": "DSSR-3D: Decoupled Reasoning for View-Dependent Referring in 3D Gaussians", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00097", "title": "DramaAgent: Agentic Storytelling Video Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00182", "title": "Localizing Post-Wire Semantic Changes in MCP Agent Frameworks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00204", "title": "Query Independent Variable Rate Visual Token Coding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00251", "title": "The Null Is the Hard Part: Exact Tests for Memorization in Generative Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00313", "title": "Rules to Tools: Executable Checks for LLM Agents in Scientific Computing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00354", "title": "Proof-Gated Signing: Solver-Checked Transaction Guards that Hold Under State Drift for Onchain AI Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00372", "title": "When Harnesses Lose the Signal: Causal Evaluation of Recovery in LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00385", "title": "FAER: Auditable Utility-Aligned Trajectory Replay for Language Model Post-Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00399", "title": "Metacognitive Reasoning in Energy Based Models using Instance Based Learning Theory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00400", "title": "Representation Transitions Reveal Emerging Safety Risks in Multi-Turn LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00431", "title": "ChainLoRA: Geometry-Preserving Task Vector Merging for Continual Learning in LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00432", "title": "XOR-Trellis: Ultra-Low-Complexity Dequantization and Curvature-Aware Hadamard-Free LLM Quantization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00435", "title": "How AI Agents Discover Scientific Equations: From Hydrotope Rediscovery to New Water-Wave Amplitudes", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00487", "title": "ScaffoldM3C: A Multimodal Sequential Monte Carlo Framework for Generative Stable Construction Planning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00497", "title": "Gumbel Straight Flow: Distilling Autoregressive Models into One-step Flow Maps", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00499", "title": "Denoising Surface: Modeling and Predicting Inference Cost for Diffusion LLM Serving", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00541", "title": "Random Recursive Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00559", "title": "PhysVista: Benchmarking Physical Intelligence in VLMs via a Perception-Reasoning-Assessment Loop", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00568", "title": "Emergent Unfaithfulness: How Alignment Training Causes Language Models to Silently Override Task Faithfulness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00573", "title": "FORTE: Adaptive Scoring and Exact Keyframe Selection for Long-Video Question Answering", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00574", "title": "Make Sparse Rewards Count: Density-Aware Reward Aggregation for Multi-Reward RL", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00604", "title": "MIKASA-Robo-VLA: Benchmarking Memory in VLA Models for Long-Horizon Manipulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00613", "title": "Spatial Strategies, Not Actions: Vector-Quantized Geodesics as Tools for LLM-Driven Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00648", "title": "Incident-Arena: Getting agents to the last nine of reliability", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00651", "title": "Agent Evaluation Reliability: More Tasks Won't (Always) Fix An Agent Leaderboard", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00661", "title": "Exploring More, Reasoning Better: Stepwise Risk-Sensitive GRPO for Diffusion Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00663", "title": "Backdoor Containment via Expert Quarantine and Shutdown in LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00666", "title": "VisionQ: VLM-as-a-Judge Taxonomy, Dataset and Benchmark for Qualitative Analysis in Computer Vision", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00675", "title": "LabBook: Harnessing Experimental History for Efficient LLM-Driven Discovery", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00710", "title": "ReLiveGym: Evaluating Long-Lived Agents over Weeks of Replayed Reality", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00717", "title": "Sequential Functional Structured Tucker Compression for Large Language Model Attentions", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00767", "title": "Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00779", "title": "Effective Synthetic Data Curation Requires Group-Level Signals", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00864", "title": "Kinematic MeanFlow: One-Step Action Generation Policy for Robotic Foundation Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00878", "title": "UniTrackPLA: Unified Panorama-Language-Action Model for Instruction-Guided Navigation and Dynamic Person Tracking", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00885", "title": "FORALL-LEAN-AGENT for Auditable Reasoning in Formal Mathematics and Software Verification", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00899", "title": "TOAST: Stochastic Robot Action Tokenization for Autoregressive Vision-Language-Action Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00952", "title": "A Matched-Budget Audit Framework for Recaptioned Image-Text Supervision Distributions", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00977", "title": "ABSENTIA: Detecting Broken Access Control Vulnerabilities in Web Applications", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00979", "title": "RISED: RubrIcs for agentic multi-environment Selection and sElf-Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00981", "title": "NarrativeFlow: Flow-Based Vision-Language-Action Model Using Robot Velocity Fields", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.00983", "title": "The Devil Is in the Reconstruction Loss Scale: Rethinking Optimization in LLM Quantization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01017", "title": "Pay for the Fault, Not the Flow: Label-Free In-Flow Multi-Agent Workflow Optimization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01023", "title": "Groundability, Not Scale Alone: When Weak Reviewers Can Audit Strong Coding Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01037", "title": "SLIM: Simplex-Lattice Interpolation Merging", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01054", "title": "Capturing In-Context Learning Dynamics with Task Operators", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01076", "title": "GLoC-EHR: Evidence-Cited Clinical Reasoning over Global Context and Local EHR Events", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01127", "title": "Counting and Min-Cost Encoding for Tokenization in Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01128", "title": "Grounding Large Language Models in DSGE Simulators for Policy Generation and Forecasting", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01172", "title": "Learning Rate Transfer for Hybrid Transformer-SSM Architectures", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01192", "title": "FlashBack: Knowing When to Remember in Streaming Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01238", "title": "Mixture-Trained Merging for Unified Multi-Objective Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01243", "title": "When the Judge Acts: Auditing VLM-Guided Image Selection on Culturally Situated Prompts", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01278", "title": "SCOPE-AD: Sequential cost-aware ordinal-belief planning with energy-based models for diagnostic agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01286", "title": "Dyna3: VLM-Guided Training-Free 4D Reconstruction via Depth Foundation Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01316", "title": "What Wins a Vote? Formatting, Length, and Lexical Diversity in the French Compar:IA LLM Arena", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01492", "title": "Q-SPT: Learnable Query-Based Compression for Low-Frame-Rate Speech Tokenization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01496", "title": "SALD: Self-Referenced Advantage Learning for Diffusion Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01499", "title": "VTR-Bench: A Systematic Benchmark for Evaluating Visual Text Rendering in Video Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01506", "title": "MCRI: A Four-Dimensional Framework for Analyzing and Evaluating Agent Skills", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01508", "title": "OverAct: Measuring and Mitigating Proactive Over-Authorization in LLM Tool-Calling Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01513", "title": "Decision Titan: Test-Time Training for Long-Term Memory in Offline Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01537", "title": "FedFit: Federated Fine-Tuning of LLMs via Vector-Bank Parameterization and Quantization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01560", "title": "AURAL: Adaptive Latent Reasoning with Joint Chunk for Speech Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01564", "title": "Chaining Skills to Hijack LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01569", "title": "Managing Context and Communication in Distributed Agentic UAV Swarms", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01595", "title": "Before It Fades: Reinforcing Temporal Representations at Inference Time in VideoLLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01605", "title": "Hob-VL: A Benchmark for Visually Grounded Boolean Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01625", "title": "Beyond Domain-Level Adaptation: Margin-Oriented Semantic-Appearance Interaction Correction for Personalized Federated Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01637", "title": "Fusing Visual and Textual Representations via Multi-layer Fusing Transformers for Vietnamese Visual Question Answering", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01640", "title": "Not All Error Yields to Scale: Where Scaling Stops in Vision-Language Inference", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01649", "title": "CrossGMN: Graph Metanetworks for Cross-Architecture Weight-Space Transformations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01670", "title": "Do MLLM Judges Judge the Edit? Auditing Bias in Image Editing Evaluation with Verified Quality Preservation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01687", "title": "Architectural Sampling: Test-Time Scaling via Computational Diversity in Frozen Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01710", "title": "CoEvolve: Construct-to-Edit Visual Grounding with Bidirectional State Refinement", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01741", "title": "ATI-VLA: Action-Centric Predictive Vision-Language-Action Models via Actionable Alignment Then Adaptive Injection", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01769", "title": "CONTRA: Discovering and Qualifying Behavior-Changing Questions for Selective Clarification in LLM Code Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01815", "title": "Debias Anything: Fairness with Diversity without Supervision in Diffusion Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01892", "title": "Selection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01917", "title": "MoLE: Mixture of Latent Experts for Complementary Visual Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01921", "title": "Cross-Lingual Alignment for Decoder-Only Models using MoE Routers", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01955", "title": "Do Your Own Research: Learning to Forecast by Learning to Search", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.01962", "title": "SIEVE: Selective attention-value Suppression for Vision-Language Models Unlearning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02005", "title": "Counting Moves, Weighing Voices: Bayesian Dialectical Argumentation for Calibrated Multi-LLM Councils under Persistent Adversaries", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02019", "title": "Controllable Multi-label Video Safety Detection via Adaptive Tversky Policy Optimization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02091", "title": "GeoLatent: Geometry-Guided Latent Structuring with Routed Optimization for 3D Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02140", "title": "Finetuning with Sampling: SFT Learns Better Than You Think", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02150", "title": "From Knowledge Access to Source Learning: Developing Source-Specific Competence", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02161", "title": "DuoMind: Enabling Distributed Multi-Robot Coordination with Semantic Communication", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02190", "title": "Trust the Direction, Search the Step: Zero-and-First-Order Methods for LLM Fine-Tuning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02191", "title": "The Missing Primitive: Diagnosing and Repairing Mathematical Reasoning in Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02196", "title": "InterEvolve: Test-Time Evolution of Reward Programs for Humanoid Loco-Manipulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02202", "title": "ScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02204", "title": "Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02206", "title": "KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02298", "title": "EditHero: A Benchmark for Long-Horizon Part-Level 3D Editing and Vibe Modeling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02304", "title": "SimuVerity: Benchmarking Agents for Engineering-Grade Simulink Model Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02323", "title": "World-Calibrated Proposal-to-Action Flow for Vision-Language-Action Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02324", "title": "Slow-Fast Multi-Teacher On-Policy Distillation for Capability Preservation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02330", "title": "Choosing Before Acting: Comparative Value Estimation for Long-Horizon Tool-Use Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02344", "title": "Drive vs. Decay: On the Training Dynamics of Joint-Embedding Predictive Architectures", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02349", "title": "MIRROR: Multipath Quorum Integrity for LLM Multi-Agent Communication", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02351", "title": "DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02359", "title": "Lexicographic Multi-Objective On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02361", "title": "SEDIMA: Cross-Run Hierarchical Insight Memory for Evolutionary Search Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02388", "title": "Octrees as an Explicit 3D Language", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02396", "title": "Inherit-MAS: Test-Time Evolution of Multi-Agent Systems through Workflow and Execution Inheritance", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02404", "title": "Trained Agentic Context Management", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02425", "title": "Finding the Move Is Not Winning the Game: XiangqiBench for Closed-Loop Evaluation of LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02444", "title": "Counterexample Generation via Per-Theorem Symbolic Verifiers: When Imitation Hurts and Reinforcement Repairs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02460", "title": "CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarking", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02462", "title": "Capability Scaling-Down Laws for LLM Compression", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02472", "title": "APDMem: Agent-Controlled Progressive Disclosure for Query-Adaptive Long-Term Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02478", "title": "Tropical Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02480", "title": "MEA: A Reward-Driven Multi-Agent System for Faithful Model Explanations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02523", "title": "Hypothesis-guided discovery of cognitive algorithms via program refinement", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02525", "title": "Learning What to Investigate Next: Meta-Reasoning for Long-Horizon Research Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02542", "title": "How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02569", "title": "Pincer: Resource Authorization for Agents using a Digital Twin", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02588", "title": "Open-Endedness Bench: Measuring Epistemic Process from Agent Records", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02593", "title": "Fisher-Guided Submodular Data Selection for Continual Pre-Training of Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02597", "title": "GRAFT: Growing Agglomerative Foundation Models via Continual Teacher Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02614", "title": "What Is Lost in Post-Training? Default Collapse and the Loss of In-Context Steerability Across Diverse Perspectives", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02616", "title": "VERSE: Verified Self-Evolving Optimizer for Agent Harnesses", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02617", "title": "WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02638", "title": "Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02657", "title": "Context-Tower Conversion Preserves Generation While Freezing Retains Knowledge: Low-Budget AR-to-Diffusion Conversion of MoE LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02666", "title": "CHASE-VLA: Post-Training Quantization Framework for Vision-Language-Action Models with Chunk-Aware Scale Estimation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02670", "title": "LEAP: Learning Efficient Action Proposals For LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02683", "title": "RAOA: Alternating-Operator Neural Computation with Programmable Radio Propagation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02687", "title": "Decoupling Memory from Context: Structured Memory for Token-Efficient Test-Time Continual Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02695", "title": "Test-time Calibration Learning for Large Language Model Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02697", "title": "GeoScaffold: Learning Compact Geometric Latents via Reconstruction for Efficient Vision-Language Navigation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02700", "title": "Learning from Evolving Errors: Adaptive Iterative Repair for On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02710", "title": "Self-Supervised Scaling of Terminal Environments for Scientific Domains", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02713", "title": "WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02718", "title": "Revisiting Visual Representation Enhancement of VLMs via Kernel Canonical Correlation Analysis", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02740", "title": "Prospective Hindsight: Self-Calibrating Reinforcement Learning via Prediction-Reality Gaps", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02781", "title": "OPD Before RL: Warm-Starting Rubric-Based RL with On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02814", "title": "VeriPy Source-Preserving Verification and Compatibility Checking for Python Components", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02828", "title": "FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02832", "title": "FastOPD: On-Policy Distillation for Lightweight VLA Deployment", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02835", "title": "All Work And No Play Makes Jack a Dull Boy: Understanding and Preventing Catastrophic Strategy Collapse in RLVR", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02840", "title": "PointWAM: 3D World Action Modeling for Dexterous Robotic Manipulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02841", "title": "How Robust Is Multimodal Claim Verification to LLM Rewriting?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02856", "title": "Adaptive Mutual Distillation for Balanced Multi-Task Post-Training of Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02876", "title": "Seeing, Saying, but Not Using: From Reportable Spatial Facts to Usable States in Multimodal Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02880", "title": "Found but Not Read: When Extracted Text Closes the Retrieval-Reading Gap in Document Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02885", "title": "PsyEvo: A Personalized Counseling Agent That Self-Evolves at Test Time", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02887", "title": "Revealing Epistemic Uncertainty in MLLMs via Causal-Invariant Masking", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02925", "title": "Positive-Unlabeled Learning for Agent Safety False Alarm Auditing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02928", "title": "Discriminating Fixture Coverage in Agent-Infrastructure Verification Suites", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02951", "title": "Dynamic Expert Pruning for Multi-Agent Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02952", "title": "GTDD: Generative Test-Driven Development for AI Coding Agents with Adversarial Testing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02970", "title": "A Guideline-Augmented Multi-Agent Framework for Schema-as-Code Biomedical Named Entity Recognition", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.02986", "title": "OLMo-Detect: A Multi-Stage, Confounder-Controlled Benchmark for Membership Inference on Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03002", "title": "Recursive Self-Improvement in Unified Multimodal Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03014", "title": "Beyond Predefined Sinks: Security-Aware Dependency Analysis for LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03020", "title": "DyadMem: A Long-Term Memory Benchmark of How Agents Work with Users", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03034", "title": "Adaptive Second-Order Solvers for Fast Stochastic Diffusion Sampling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "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", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03039", "title": "HyperThink: Text-to-Parameter Hypernetworks for Efficient Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03063", "title": "HARPO: Hallucination-Aware Reinforcement Learning for Faithful and Creative Language Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03084", "title": "NegT2IBench: When Negation Changes the Picture. A Polarity Benchmark for Text-to-Image Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03089", "title": "Securing Computer-Use Agents Against Branch Steering Attacks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03099", "title": "Beyond Single Videos: Benchmarking and Active Evidence Seeking for E-Commerce Cross-Video Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03102", "title": "Ask, Relax, or Act? Evaluating Actionable Indeterminacy in LLM Preference Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03123", "title": "Foresight: planning future perception in streaming VLMs without retraining", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03190", "title": "Not Until the Evidence Says So: Teaching LLM Investigators When to Close a Case", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03198", "title": "KV$^2$: A Self-Refining KV Cache", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03199", "title": "Predicting and Repairing Merge Collapse in Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03213", "title": "Toward SLM-based agentic task-tool intent matching", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03226", "title": "D2K-Bench: Can LLM Agents Turn Expert Designs into Efficient GPU Kernels?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03315", "title": "Lightweight, Rubric-Guided Trajectory Evaluation for Production AI Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03389", "title": "From Patching to Pruning Visual Computation in Vision Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03448", "title": "Passing the Test You Trained On: Re-evaluating Prompt-Injection Detectors for LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03516", "title": "XGenAct: Geometry-Enhanced World Action Models through Cross-Task Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03524", "title": "From Benchmarks to Production: A Text-to-SQL System for Complex Financial Data", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03574", "title": "HyperBrowseComp: A Multilingual and Multimodal Stress Test for Web-Browsing Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03585", "title": "Threat-Preserving Representation Sensitivity in Agent-Security Benchmarks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03591", "title": "HazardWeaver: Scientific Route Selection for Hazard Analysis Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03631", "title": "NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03632", "title": "World Embedding Benchmark", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03695", "title": "Language Models that Play Chess and Explain Their Moves", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03741", "title": "Logit-Aware MIMO AirComp for Distributed Mixture-of-Experts LLM Inference over Wireless Edge Networks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03780", "title": "OceanMind: A multi-agent AI system for ocean diagnosis", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03796", "title": "ProsaBuddy: Assisting Mechanized Real-Time Schedulability Analysis with LLM-based Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03820", "title": "From Requirements to Attack Trees: Grounded LLM Agents for Design-Time Security Review", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03894", "title": "Proxy Confidence: Auditing Black-Box LLM Agents with a Surrogate's Log-Probabilities", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03938", "title": "MLLMs Fail to Refuse when Using Tools Agentically", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03984", "title": "Teaching Agents to Code Reliably", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.03994", "title": "Solving VeriContest with a Lean-Backed Rust Verifier", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04002", "title": "Do Language Models Need a Trainable Input Embedding Table? Fixed Minimal Token Codes at 1.7B-Class Scale", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04008", "title": "SkillScriptBench: Benchmarking Self-Evolution of Executable Agent Skill Packages Beyond Markdown", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04009", "title": "SUAVE: Unified Video-Action Models via Masked Diffusion", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04047", "title": "Periscope: Extending Frozen Language Models Beyond Their Context Window", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04118", "title": "LongSocialBench: Do Long-Context LLMs Understand Online Discussion Threads?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04132", "title": "PB-GRPO: Learning Socially Adaptive LLM Agents from Persona-Driven Simulation with Preference-Batched GRPO", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04134", "title": "Agentic Resource Allocation for Batch Multi-Objective Bayesian Optimization in Autonomous Materials Discovery", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04139", "title": "From Sight to Foresight: Predictive Spatial Reasoning in Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04140", "title": "ExpertMuon-Compass: Alignment-Guided Step Sizes for Mixture-of-Experts Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04152", "title": "Kepler4D: Controllable Future Video Generation via 4D Scene State Evolution", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04178", "title": "SHarP: Saliency-based Pruning of Agent Harnesses", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04196", "title": "Asynchronous Is Nearly Free for Evolution Strategies on Long-Horizon Agentic Tasks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04204", "title": "Clean: Second-order LLM Training at Linear Memory Cost via Nyström Sketching", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04225", "title": "FlashGaze: Training-Free Multi-Scale Patch Pruning For Efficient Video Understanding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04261", "title": "Playing social deduction games with reinforcement fine-tuned large language models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04292", "title": "LMBuild: Evaluating LLM Agents for Generating Buildable and Functional Structures", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04299", "title": "Questioning the Questions: Sustaining Self-Evolution in Reasoning Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04303", "title": "What to Preserve in Recursive Computation: A Local Predictive Sufficiency Principle", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04318", "title": "Rethinking Long-Video Efficiency: A Joint Allocation Perspective on Frames, Pixels, and Front-End Latency", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04328", "title": "Bidirectional Preference Synthesis: Learning Prompt-Conditioned Preferences from Boundary Failures", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04344", "title": "Hierarchical Credit Assignment for RLVR on Fused Gromov-Wasserstein Geometry", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04371", "title": "Functionally Equivalent or Not? Graph-Grounded Differential Surrogate Execution for Code Equivalence", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04375", "title": "Do Tool Calls Execute as Intended? Measuring and Repairing Intent-Execution Correspondence in LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04379", "title": "AgentPersonaBench: Benchmarking Persona-Driven User Simulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04381", "title": "Beyond Plausibility: Verifiable Fine-Grained Image Editing on Structured Assets", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04409", "title": "Understanding and Mitigating Hallucination Escape in Tool-Using LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04429", "title": "Asking Earns Nothing: Scoring the Decision to Act in BFCL Multi-Turn", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04432", "title": "Video2World: Benchmarking Coding Agents for Interactive World Modeling from Embodied Videos", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04488", "title": "Can LLM Agents Automate Reinforcement Learning for Text-to-Speech?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04506", "title": "EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04512", "title": "Correctness Is a Direction: Geometric Answer Selection in Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04517", "title": "EvoCast: Reliable Autonomous Research Agents for Iterative Forecasting Architecture Evolution", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04518", "title": "Length Generalization Needs Proper Regularization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04555", "title": "$\\mathrm{TRIZ}^{a}$: Guiding Agent Evolution from Pattern Recognition to Solution Invention", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04575", "title": "From Probe Scores to Alarm Policies: Operational Validity of Activation Monitors for Language-Model Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04580", "title": "Strong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04596", "title": "DiffGate: Difficulty-Gated Teacher Guidance for On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04604", "title": "Anticipating the Consequences of Curriculum Decisions with Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04607", "title": "ForeAct3D: Policy-Grounded Future World Modeling for VLA Policies", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04611", "title": "Practical Runtime Enforcement of First-Order Temporal Requirements", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04616", "title": "PerturBot: Breaking Shortcut Priors in Vision-Language-Action Models with Perturbative Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04641", "title": "Back to the Future: Regressing Readability Features from LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04646", "title": "SEIS: Self-Evolving Inference Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04651", "title": "GS-Codec: A Gaussian-Splatting Bottleneck for Neural Audio Coding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04658", "title": "RETRACE: From Entangled Repair Histories to Reusable Experience for CI Repair", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04687", "title": "Understanding Errors in LLM-Based Question Answering over Imperfect Tables", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04699", "title": "Not Self-Decidable: LLMs Cannot Draw the Boundary of What an Agent Verifier Can Check", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04721", "title": "Knossos and Ariadne: Benchmarking and Learning Complete Diagram Topology Extraction with Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04753", "title": "More Value per Key: Asymmetric Sparse Attention for Faster LLM Decoding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04779", "title": "Asymmetric Repository Lineage Modeling and Verifier-Guided Coordination in Concurrent AI Coding Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04792", "title": "Do More Modalities Always Help? A Geometric Perspective on Missing-Modality Robustness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04794", "title": "Knowing the Store: What a Memory Backend Must Write Down Before an Agent Can Read It", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04805", "title": "ExStereo: Lifting 2D Vision-Language-Action Models to 3D with Explicit Stereo Representations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04824", "title": "Agent Behavior as Code: Efficient and Robust LLM Agents with Programmatic Specifications", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04845", "title": "Which Preferences to Train On? End-to-End Multi-Objective Alignment with an Adversarial Preference Distribution", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04851", "title": "CURIO: Curiosity-Driven Test-Time Learning for Open-Ended Discovery", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04868", "title": "From Memory to Guide: Spatio-Temporal Composer for Procedural Coding Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04882", "title": "ResOPD: Tail Residualization for Sparse On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04889", "title": "ForkPilot: Self-Evolving Policy for Retrospective Search in Long-Horizon Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04899", "title": "Rewrite What Matters: Adaptive Multilingual Query Rewriting for Reasoning via Agentic Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04915", "title": "Are We Measuring Scientific Intelligence? Rethinking the Evaluation of AI Scientists", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04916", "title": "PreAct-Nav: Agentic Reasoning Before Action for Urban Navigation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04918", "title": "Residual Visual Credit Optimization: Conserved Evidence Routing for Multimodal Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04921", "title": "Complex Agents, Shallow Tests: Demystifying and Enhancing Test Adequacy of Agent Harness in the Wild", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04927", "title": "Assembling Insights for Agentic Machine Learning Engineering Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04940", "title": "Software World Models: From Consequence Prediction to Decision Value", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04950", "title": "How Should Teachers Be Prepared? RL on Student-Induced States for On-Policy Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04961", "title": "Building LLM Agent Systems the Deep Learning Way: From Modular Design to Architecture Search", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.04991", "title": "CIPO: Counterfactual Imagination Policy Optimization for Adaptive Tool Granularity Selection", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05023", "title": "Look Where You Say You're Looking: Self-Grounded Attention for Visual Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05024", "title": "LightVLN: Efficient Aerial Vision-and-Language Navigation with Compact Memory and History-Guided Local Aggregation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05025", "title": "Triggering Generalist Reasoning via Predictive Uncertainty for Dual-System VLA", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05026", "title": "GeoBridge-VLA: Geometry-Aware Residual Adaptation for Vision-Language-Action Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05030", "title": "EVISKILL: Grounding Skill Evolution in Replayable Evidence", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05039", "title": "Causal Improvement Graph for Agentic Harness Optimization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05047", "title": "Beyond Task Completion: Measuring Interaction Cost in Terminal User Interfaces", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05066", "title": "Salvation Lies Within: Eliciting Inherent Style Transfer in Step-Distilled Diffusion Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05097", "title": "ReMAP: Restoring the Perceptual Cycle with Reasoning-Time Latent Visual Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05099", "title": "Small Agents with Semantic Search: Efficient Multilingual Code Localization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05107", "title": "SearchJev: A Fast and Calibrated System-1 Model for Search Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05124", "title": "Memory Canonicalization: A Framework and Benchmark for Cross-Model Drift in Persistent LLM Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05126", "title": "Selecting Repetition Counts Across Model Scales in Data-Constrained Pretraining", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05132", "title": "SpecAgent: Empowering Program Verification with Agentic Synthesis of Formal Program Specifications", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05139", "title": "Hidden in the Comments: A Context-Injection Attack Surface in Code LLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05162", "title": "Memadapter: Counterfactual Adaptation Against Memory-induced Sycophancy", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05176", "title": "AECG: Asymmetric Experience Consolidation and Governance In Multi-Agent Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05185", "title": "Recurrent Latent Visual Search for GUI Grounding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05190", "title": "FORGE: Verification-Gated Behavioral Repair for Generative Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05247", "title": "templar: agentic induction and evolution of standardized radiology reporting templates from large-scale clinical corpora", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05273", "title": "When and What to Prune? Stage-Aware Visual Token Pruning for Efficient VLA", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05274", "title": "Answer with Evidence: Consistency-Aware Grounded Visual Question Answering for Roadside Traffic Scenes", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05282", "title": "Red-TTT: Test-Time Training for Automated Jailbreaking Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05284", "title": "Fusion is the New Mutation: Bandit-Guided Evolution on Workflow Graphs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05295", "title": "Readable Before Actionable: Causal Tracing of Indirect Prompt Injection", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05300", "title": "MESH-Harness: Self-Improving Agent Harnesses via Bandit-Guided Compositional Evolution", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05303", "title": "ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05308", "title": "RubricArmor: Adversarial Evolution Improves LLM-Based Rubric Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05318", "title": "Robust Parameter-Efficient LLM Adaptation on Analog Hardware", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05334", "title": "AgentDiscover: Autonomous Discovery with Minimal Search Scaffolding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05342", "title": "IRSTD-Agent: Agentic Infrared Small Target Detection via Zoom-Guided Interaction Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05367", "title": "AIProver: Agentic Auto-Formalization of Mathematical Research via Certificate-Driven Evolving Harness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05373", "title": "Towards Unbiased On-Policy Distillation for Block Diffusion Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05383", "title": "Sibyl: An Efficient Small-large Model Collaboration Framework for Long-horizon Tasks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05398", "title": "MMPostTrainBench: Benchmarking Autonomous Research for Multimodal Post-Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05402", "title": "Task Vector Descent: Learning from Non-IID Batches", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05418", "title": "EvoMem-VLA: State-Evolution Memory for Long-Horizon Robot Manipulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05425", "title": "Unmentioned Checklist Findings Change How Reinforcement Learning Appears to Improve Chest Radiograph Report Checking", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05473", "title": "Hierarchical Reinforcement Learning with Stable Temporal Abstraction for Language Model Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05481", "title": "CodeForge-MA: Execution-Verified Multi-Agent Learning with Language-Conditioned LoRA for Multilingual Code Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05484", "title": "Universal Test-Time Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05532", "title": "DelegationBench: Measuring When AI Agents Should Ask Before Acting", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05538", "title": "LiFT: Loop Flow Transformers", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05540", "title": "SALUS: Automated Auditing of NL-to-SQL Benchmarks through Weak Supervision of Multi-Agent Output", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05563", "title": "G-CARB: Graph-Localized Conformal Agent Risk Budget for Compositional Harm", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05572", "title": "Disentangling Task Difficulty from Run-Level Failure in Agent Failure Prediction", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05615", "title": "DREAM: Dynamic Resolution Assignment For Multimodal Multi-agent Debate", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05622", "title": "UndoBench: Separating Task Competence from Recovery Capability in Tool-Using AI Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05630", "title": "Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05637", "title": "Visual Grounding Safety in Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05664", "title": "StageVLN: Spatial and Trajectory Auxiliary Guidance for Efficient Vision-Language Navigation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05685", "title": "Spend Bytes on Breadth: Precision-Count Trade-offs for Decode-Time KV Compression in Long Chain-of-Thought Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05697", "title": "From Token-Max to Outcome-Max: How You Use AI Determines Its Productivity", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05709", "title": "Nexus: An Execution Fabric for AI Agents Across Cloud, Edge, and Devices", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05715", "title": "Rotated, but How Far? Diagnosing and Improving Object-Rotation Reasoning in VLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05744", "title": "CIPHER-MoE: Balancing Efficiency and Routing Fidelity in Trillion-Scale MoE Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05774", "title": "AdaSpark: Adaptive DSpark with Online Learning for Tree Verification and N-gram Fill", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05778", "title": "MedicalHarness: A Controlled Evaluation of LLMs and Agent Harnesses on Medical Tasks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05833", "title": "Request Order Matters: Cache-History Sensitivity in Selective KV-Cache Reuse for Rolling Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05842", "title": "HLA: Expressive Hybrid Linear Attention via Chunk-Wise Dynamic Mixing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05861", "title": "Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05872", "title": "Off-Policy Merging Beats On-Policy Self-Distillation for Continual Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05879", "title": "Learning to Learn a Language", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05885", "title": "The Optimization Landscape of Learning Compacted Context Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05896", "title": "TasteRoute: Personalized Routing for Video Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05897", "title": "Fitting Vision Adapters at Frontier Scales", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05935", "title": "ThunderSyncRL: Lossless Acceleration of Agentic Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05940", "title": "ReMem: Streaming Video Understanding With Long Context Retention", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05978", "title": "Byte Language Models: Scaling, Emergent Abstractions, and Information Allocation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "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", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05993", "title": "From Transformation to Target State: Rethinking Query Representation for Zero-Shot Composed Image Retrieval", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.05994", "title": "How (and How Not) to Use Data Augmentation in VLA Post-Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06001", "title": "AgentSpy: Making AI Agent Behavior Observable", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06021", "title": "Scalable Minimal-Change Learning for Controllable Image Editing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06026", "title": "Differentiable Bit-Widths: Co-optimizing Pruning and Quantization via SVD for Ultra-Efficient LLM Compression", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06064", "title": "TrustMI: Causally controlling how assistants trust their users", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06100", "title": "From Traces to Agentic Worlds: Agentic Language World Models for Interactive Environment Simulation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06147", "title": "Efficient Test-time Adaptation through Candidate Verification and Divergence Shifts", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06161", "title": "Introducing Code-Switched Contexts to Cognitively-Inspired Bilingual Model Training", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06191", "title": "Judged Useless, Queried Anyway: Tool-Using Agents Rarely Turn Their Own Evidence Judgments into Stopping Decisions", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06192", "title": "Copies or Sources? Measuring How LLM Aggregators Count Restated Evidence in Multi-Agent Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06193", "title": "Correct Code, Broken Contributions? SWE-CC: Benchmarking Repository Policy Compliance for Coding Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06226", "title": "LeAVJEPA: A Minimalist Architecture for Audio-Visual Self-Supervised Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06269", "title": "Evolving in Thought Space: Training a Small Model at Test Time Unlocks Better Discoveries", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06286", "title": "DeferKV: Rethinking Eviction Timing for One-Shot KV Cache Compression", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06324", "title": "Readout Blindness: VLM Scores Miss the Spatial Direction Their Frozen Encoders Retain", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06404", "title": "Quantifying the Stability of Multi-Step Reasoning via Error Amplification", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06406", "title": "What Did the Agent Actually Do? Evidence-Grounded Oversight for Long-Horizon Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06413", "title": "SpatialChain: A Benchmark for Auditing Spatial Reasoning Faithfulness in VLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06439", "title": "Better Call Reward: Reward Hacking as Strategic Abstention in Legal Reasoning Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06446", "title": "The Assistance Dilemma: Learning to Teach via Multi-Turn Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06454", "title": "AgentPrivArena: Evaluating and Auditing Real-world AI Agent Privacy", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06479", "title": "Behavior-Preserving KV Cache Compression", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06598", "title": "SimForcing: Distilling Simulation Motion Priors into Real-Domain Robot World Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06611", "title": "Lens3D: Target-Conditioned Visual Foveation for Fine-Grained 3D Understanding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06616", "title": "Video Encoders Built on Image Representations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06617", "title": "RealtimeWAM: One-Step Asynchronous World Action Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06636", "title": "Differentially Private Mixing of Public Datasets Improves Private Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06637", "title": "Long-Horizon Textual World Modeling through Structured Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06647", "title": "LoGRA: Scaling LLM Reinforcement Learning with Low-Rank Gradient Sketches", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06648", "title": "Representation-Space MMD for Diffusion Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06650", "title": "Wikidata Search Traces: A Dataset for Training Knowledge Graph Search Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06666", "title": "What Matters for Latent Reasoning with Flow Matching", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06671", "title": "Learning What to Imitate: Entropy-Aware Distribution Mixing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06686", "title": "OVAL: Output-Aware Local Page Bases for KV Cache Retrieval", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06729", "title": "Improving Diversity in LLM Short Story Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06748", "title": "BazaarBench: Delegation Safety in Decentralized C2C Marketplaces Run by LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06750", "title": "Balancing Memory Pathways: Analyzing and Improving Memory Utilization in Hybrid LMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06751", "title": "MatrixFormer: A Foundation Model for Matrix Completion", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06782", "title": "T-Search: An Open Agentic Retriever and Playground for Hard Multi-Step Search", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06804", "title": "Sharpen Without Search: On-Policy Distillation of Sequence-Level Power Distribution", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06817", "title": "Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06829", "title": "CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06830", "title": "MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06843", "title": "Recursive Video In-Context Learning for Agentic Robot", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06844", "title": "Learning to Read the Contextual Tokens in Diffusion Transformers", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06892", "title": "Axiom Satisfiability of Linear Rewards in Alignment", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06932", "title": "RADC: Risk-Aware Dual Caching for Vision-Language Test-Time Adaptation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06947", "title": "FactorBench: A Portfolio-Aware Benchmark for Automated Factor Mining", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06950", "title": "Learning to Decide, Not to Reason: Parameter-Efficient Decision Operators via Low-Rank Activation Steering", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06963", "title": "WavePrune: One period is often enough for RoPE", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06964", "title": "Principles that Guide, Actions that Inform: Agent Evolution via Knowledge Abstraction", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06966", "title": "APEX: Active Protection at Execution Boundaries for LLM Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.06993", "title": "DART-ES: Difficulty-Aware Reweighting and Targeted Replay for Fine-Tuning LLMs with Evolution Strategies", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07043", "title": "TRIAGE: Direction-Aware Mismatch Stabilization of Native NVFP4 Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07118", "title": "AMBER: Training Long-Horizon Web Agents through Append-Only Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07121", "title": "SoloQ: Calibration-Free Quantization for Diffusion Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07132", "title": "CroissantMiner: Automated Extraction and Validation of Croissant Metadata for ML Datasets", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07177", "title": "CLM-as-a-Judge: Evaluating an Open Contrastive Decision Model on Public Judge Benchmarks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07191", "title": "Agentic Design Space Exploration for Joint Hardware Configuration Selection and Mapping of AI Inference Workloads on Heterogeneous Edge SoCs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07212", "title": "Reward-Driven Learning under Prompt-Level Differential Privacy", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07247", "title": "Learning What to Distill: Bilevel Top-K Token Selection for Self-Distillation in Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07258", "title": "Lineage-Aware Memory Governance: A Derivation-Gated Framework for Privacy-Preserving Column-Level Access Control in Enterprise AI Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07261", "title": "Verifying Coordination in Parallel Coding Agents: NP-Bench and a Scheduling Planner", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07289", "title": "Catching Developers in the Flow: Low-Latency Agentic Program Repair at Google Scale", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07311", "title": "Understanding and Mitigating Inference-Time Overreliance Using Agentic Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07324", "title": "Scale-Invariant Training for Time Series Foundation Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07332", "title": "Structuring MoE Expert Selection for Agentic Reinforcement Learning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07335", "title": "Selective Critique for Cost-Aware LLM Agents in Long-Horizon Decision Making", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07348", "title": "Stepped MoE: Segment-Level Routing with Configurable Inference Complexity", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07349", "title": "RELACE: retrospective likelihood-based action credit estimation for long-horizon language agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07376", "title": "MemCo: Memory-Centric Collaboration for Generalizing LLM Agents to Unseen Environments", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07385", "title": "Semantic Capability Acquisition and Specialization During Vision-Language Model Fine-Tuning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07389", "title": "Inference and learning in sparse autoencoders as natural gradient flow", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07457", "title": "AlignQuant: Tile-Aligned Mixed-Precision Quantization for Efficient LLM Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07522", "title": "Activation Denoising: A Robustness View on Parallel vs Sequential LLM Quantization", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07532", "title": "Safeguarding LLMs via Model-Agnostic Latent Safety Signals from Dark Knowledge", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07533", "title": "SkillFormer: Skill-Decomposed Adaptation for Audio Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07557", "title": "CheckerBench: Can Long-Horizon Agents Synthesize Static-Analysis Checkers?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07572", "title": "Two Vectors Replace In-Context Demos: Structured Task Adaptation via Embeddings", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07585", "title": "REViT-v2: Hierarchical Windowed Roto-reflection Equivariant ViT for Equivariant Feature Extraction", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07587", "title": "Large Language Model Orchestration under Heterogeneous Preferences via Explicit Persona Inference", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07639", "title": "HarnessSecurity-Bench: Do Security Mechanisms Really Protect Coding Agent Harnesses?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07643", "title": "Monte Carlo Estimation for KV Cache Eviction", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07645", "title": "SkillPoison: Progressive Skill Poisoning via Successful Experiences", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07654", "title": "Does On-Policy Distillation for Safety Pose Backdoor Risks?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07657", "title": "Where Rules End and Judges Begin: Measuring the Judgment Boundary in Multi-Agent Systems Security", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07689", "title": "Unlocking Fine-Grained Perception in CLIP via Structurally-Aware Latent Masked Modeling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07706", "title": "WASD: Wasserstein-based Knowledge Distillation for Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07729", "title": "Foveated Compression: Selective High-Resolution Preservation for Token-Efficient VLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07731", "title": "Learning to Retrieve via Reinforcement Learning in Embedding Space", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07739", "title": "Cite What You Explore: Budget-Aware LLM Reasoning over Medical KGs with Verifiable Evidence", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07758", "title": "Later Is Better: Token Reduction for ViTs Under Distribution Shift", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07767", "title": "TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07774", "title": "Reading, Not Manipulating: Leveraging Router Logits for Multimodal Safety in MoE Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07809", "title": "MASKerade: Token-Routed Mask Experts for Dense-to-MoE Upcycling", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07810", "title": "SIFT: Search Intent-to-Filter Transformer for Multi-Task Personalized Filter Ranking at Airbnb", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07816", "title": "Do I Need the Cloud? Uncertainty-Aware Step-Level Handoff for Small Language Model Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07819", "title": "$α$Transfer: Coefficient Transfer for Efficient Model Merging", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07832", "title": "Harness Engineering for Software Engineering via Modular Executable Dev-Primitives", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07848", "title": "Dynamic Positional Attention Modulation for Parameter-Efficient Fine-Tuning of Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07851", "title": "RA-MoWE: Workflow-Affinity Embeddings for Query Clustering and Agentic Workflow Generation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07860", "title": "WorkflowOps: Learning Agent Collaboration Priors for Multi-Agent Workflow Orchestration", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07862", "title": "A self-learning scientific agent for X-ray diffraction", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07868", "title": "CueRator: Agentic Search for Symbolic Rules to Adapt Frozen Multimodal Encoders", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07887", "title": "Visual Abstention in Unified Multimodal Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07898", "title": "FC-SWE: Failure-Conditioned RL for Long-Horizon Software Engineering Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07903", "title": "Unsupervised Long-Tailed Adaptation of Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07925", "title": "TF-PRVR: Training-Free Partially Relevant Video Retrieval", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07935", "title": "SIGMA: Self-Improving Alignment Generalization from a Model Spec", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07940", "title": "Hybrid Latent Attention for Looped Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07972", "title": "Can Agents Work for Everyone? Cross-User Reliability for Mobile GUI Agents in Personalized User Interfaces", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07987", "title": "VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.07990", "title": "A Broader Look at Model Merging: Rethinking Implicit Regularization Induced by Task Arithmetic", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08048", "title": "DAEDALUS: Bootstrapping Agent Memory from Self-Generated Tasks", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08076", "title": "SpeedrunBench: Challenging LLM Agents with Video Game Speedrunning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08101", "title": "Beyond Corrected Memory: Execution Consistency in Multi-Agent Systems", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08106", "title": "ChartBmkAgent: Harness-Governed Multi-Agent Construction of Chart QA Benchmarks from Sparse Error-Taxonomy Specifications", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08108", "title": "Enhancing Diffusion Language Models with Autoregressive Post-Training Weights", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08133", "title": "VLA-ACL: Action-Consistent Visual Token Pruning for Efficient Vision-Language-Action Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08155", "title": "Token-Efficient Multi-Agent Collaboration via System One-Guided Computational Division of Labor", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08159", "title": "Making COMET Comparable Across Scripts: Diagnosis and Correction of Tokeniser-Induced Script Bias in Indic MT Evaluation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08162", "title": "The Failure Is in the Readout: Fine-Grained Emotion Recognition Benchmarks Measure Elicitation, Not Perception", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08164", "title": "Align, Then Correct: Training-Free Two-Stage Low-Rank Compensation for Extremely Quantized Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08183", "title": "Compact Robot Policies Need Fine-Grained Visual Representations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08341", "title": "DIPrune: Task-Aware Token Pruning with Dual Importance for Efficient Multimodal Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08364", "title": "Transect: Retaining Observability for Long-Horizon LLM Agent Evaluations", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08401", "title": "GeoPID: Decomposing and Steering Visual Information in Vision-Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08403", "title": "SSR: Sparse Segment Reduction for Ternary GEMM Acceleration", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08414", "title": "Image Bitstream Fine-grained Understanding for Privacy-Friendly AIoT", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08452", "title": "Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08463", "title": "UNREAL: Unifying Retrieval and Long-Context with a Single Model", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08514", "title": "How Much Evidence Should a Coding Agent's Self-Correction Carry? Adaptive Dirichlet Evidence for Self-Distillation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08560", "title": "Have I Seen Enough? Frozen Video-Language Models Encode Evidence Readiness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08586", "title": "MINDSET: Energy-based Schema Evolution for Long Conversational Agent Memory", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08621", "title": "Recursive Game Creator: An Agentic Product-Level Experience-Oriented Game Harness", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08622", "title": "Agentic RCA for Internet-Scale Services Using Constrained Creativity", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08639", "title": "Forensic Reserve: Eliciting Latent Knowledge for Image Forgery Detection", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08647", "title": "SquidAgent: Parallelize Wisely, Coordinate Efficiently", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08649", "title": "Stable Scores, Unstable Answers: Frame Phase and Option Order in Video Multiple-Choice Evaluation", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08659", "title": "Selective Transfer of RL Updates for Visual Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08662", "title": "ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08670", "title": "Principled Under Pressure: Post-Training Decides Whether LLMs Act on Their Own Moral Judgment", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08674", "title": "EC-RAG: Event Chain Retrieval-Augmented Generation for Long Video Understanding", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08678", "title": "Secure Speculative Decoding for Large Language Models", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08691", "title": "ScienceClaw: Benchmarking Continual Self-Evolution of AI-for-Science Agents Across the Natural and Social Sciences", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08713", "title": "SpaTime: Streaming Vision-Language Models for Spatio-temporal Reasoning", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08718", "title": "When Forgetting is not Catastrophic: On the Mechanics of Spurious Forgetting", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08773", "title": "AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08775", "title": "Agent in a Bottle: Can LLM Agents Turn Their Capabilities Into Cheap, Scalable Artifacts?", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08778", "title": "Sherpa: Teaching LLMs to Teach Adaptively", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08779", "title": "ALIVE: Interaction-Aligned Object Insertion for First-Frame-Guided Video Editing", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.08791", "title": "World Models' Last Exam in Physics", "old_status": "implementation-pending", "accepted_registers": [ "oct07", "oct12" ] }, { "id": "2610.18157", "title": "Fan Out or Relay? Ownership-Aware Routing for LLM Agents Querying Siloed Tables", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.18684", "title": "Knowing the User Is Not Knowing What We Established: Grounding Awareness in Personal Agents", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.18780", "title": "BLADE: Auditing Information Use for Failure Attribution in LLM Multi-Agent Systems", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19131", "title": "Plant Memory: Technical Language Processing and a Verified Maintenance Ontology for Agent-Accessible Failure Knowledge from SAP PM Work Orders", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19139", "title": "Online Tokenization with Optimal Delay in Linear Time", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19316", "title": "Truthful and Expected-Payment-Optimal Model Routing in Graph-Based Agentic Workflows", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19345", "title": "When Is Smaller Greener? Right-Sizing Language Models for Agentic AI", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19474", "title": "Beyond Retrieval: Evidence Access and Event Aggregation in Agent Memory", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19659", "title": "Measuring and Reducing Unusable Replies of LLM Agents", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19801", "title": "Delta Tokenizer: Reference-Identical Tokenization of Session Deltas and Segment-Parallelism for Long Prompts on Multicore CPUs", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.19818", "title": "One Monitor Does Not Fit All: Formalism-Agnostic Runtime Verification for Agentic AI", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.20067", "title": "Which Knowledge Should LLMs Trust? Uncovering a Generalizable Steering Direction Across Knowledge Conflicts", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.20303", "title": "SpectrumQuant: Spectrum Alignment for LLM Quantization", "old_status": "reviewed", "accepted_registers": [ "oct15" ] }, { "id": "2610.20413", "title": "When Does Momentum Help in Reinforcement Learning with Verifiable Rewards?", "old_status": "reviewed", "accepted_registers": [ "oct15" ] } ], "missed14_details": [ { "id": "2610.18157", "title": "Fan Out or Relay? Ownership-Aware Routing for LLM Agents Querying Siloed Tables", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "OAR 仅用公开schema生成一份草拟SQL并计算owner cover，在读任何私有行前决定fan-out或relay。核心可在小型真实数据库和实际SQL草拟模型上低成本检验；不能直接喂gold SQL/owner答案替代草拟，不声称13种设置已复现。" }, { "id": "2610.18684", "title": "Knowing the User Is Not Knowing What We Established: Grounding Awareness in Personal Agents", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "Grounding Layer 明确分开用户事实与已接受、未确认、修订和限域约定；ledger、动作前gate及回查原始对话均保留。可低成本核验真实对话到状态提取及有限动作边界；不得直接喂入oracle +Status/+Policy或gold动作当实现。只接受机制候选，不声称全基准收益。" }, { "id": "2610.18780", "title": "BLADE: Auditing Information Use for Failure Attribution in LLM Multi-Agent Systems", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "BLADE 审计来源到决策的使用关系，区分引入、传播和修复；原文明确不输入参考答案。全文片段报告1228轨迹的 Jev 推断估算约$1.9，属于廉价真实轨迹审查候选；仍需真实决策模型，不得用金标责任步骤做 fixture。" }, { "id": "2610.19131", "title": "Plant Memory: Technical Language Processing and a Verified Maintenance Ontology for Agent-Accessible Failure Knowledge from SAP PM Work Orders", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "Plant Memory 的 rule-first TLP、带证据/审核状态的事实图及 SQLite 离线后端可低成本实现，LLM第二读者是可选项。接受规则/溯源核心候选；400条合成工单的数字只是受控机制证据，不能宣称真实工厂部署、通用Agent能力或自动evolve收益。" }, { "id": "2610.19139", "title": "Online Tokenization with Optimal Delay in Linear Time", "batch": "2026-10-15-review-closure-foundation-model", "resolution": "affordable_candidate_not_implemented", "reason": "在线tokenization以稳定输出前缀与delay形式化为基础，组合special-token、pretokenizer和BPE，并给出更低delay的联合算法与不可解边界；公开Morgan Stanley上游实现可做CPU级精确ID/流式等价性验收。接受为低成本算法候选，本轮不实现、不跑完整吞吐研究。" }, { "id": "2610.19316", "title": "Truthful and Expected-Payment-Optimal Model Routing in Graph-Based Agentic Workflows", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "STAR 的分配目标、工作量单调性及积分支付可在小型真实合成图上用精确求解器执行，不需要调用 LLM。接受的是拍卖/路由核心机制候选，不把合成成本或激励保证当作真实 Agent 能力提升。" }, { "id": "2610.19345", "title": "When Is Smaller Greener? Right-Sizing Language Models for Agentic AI", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "本文的贡献是按成功完成任务计量能耗/成本/时间的决策公式与Pareto筛选。接受CPU数值机制候选，可从公开参数算交叉长度；15模型GPU能耗pilot并非已复现，也不能将合成可靠率下的公式结果宣传为Agent能力。" }, { "id": "2610.19474", "title": "Beyond Retrieval: Evidence Access and Event Aggregation in Agent Memory", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "TallyMem 保留原始轨迹、生成事件规格并实际执行聚合，不能以金标事件集代替规格生成。接受原始日志→事件ledger→真实编译/执行的廉价核心候选；$1.05仅属原文primary arm，不能外推全AMA/多reader成本。AMA混入开发轨迹，正式能力比较应采用25开发/99隔离仓库协议。" }, { "id": "2610.19659", "title": "Measuring and Reducing Unusable Replies of LLM Agents", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "回复可用性把读者实际见过的信息作为边界，单次重写或追加明确回复要求无需训练。接受少量真实回复上的reader检查与实际重写机制候选；不能用金标可用回复替代重写器，480回复/8模型统计不视为本地结果。" }, { "id": "2610.19801", "title": "Delta Tokenizer: Reference-Identical Tokenization of Session Deltas and Segment-Parallelism for Long Prompts on Multicore CPUs", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "Delta Tokenizer 的核心是由 tokenizer 拆分规则证明安全的增量接缝、精确分段及 chunk 调度，而不是简单拼接 token。CPU 核心可在本地 Mac 以真实 tokenizer 和原始文本核验一致性；不复跑 41.8B 文档或多服务器服务实验，不声称已测 TTFT。" }, { "id": "2610.19818", "title": "One Monitor Does Not Fit All: Formalism-Agnostic Runtime Verification for Agentic AI", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "AgenticRV 的 canonical events、monitor-specific projection、统一 verdict 与 effect 前阻断是可执行的运行时机制。Python/受控工作流可在 Mac 保真核验；只接受该机制，不把立即返回的 monitor 微基准冒充语义监控效果或 Agent 能力。" }, { "id": "2610.20067", "title": "Which Knowledge Should LLMs Trust? Uncovering a Generalizable Steering Direction Across Knowledge Conflicts", "batch": "2026-10-15-review-closure-agent", "resolution": "affordable_core_review_candidate", "reason": "知识冲突偏好方向是残差空间的线性干预，原文采用7B–9B、单卡A100、仅推理。接受真实模型上独立开发冲突对提取方向与隔离测试的核心候选；不把先按clean正确筛出的条件准确率当无条件能力，也不把评测答案送入推理。" }, { "id": "2610.20303", "title": "SpectrumQuant: Spectrum Alignment for LLM Quantization", "batch": "2026-10-15-review-closure-post-training", "resolution": "affordable_candidate_not_implemented", "reason": "SpectrumQuant 用校准激活协方差构造闭式 alignment+mixing 正交旋转；原文7B校准约0.26 GPU小时、4.51 GiB，可保留128条真实校准样本的7B原算法路径；不承诺70B或全套多模型下游扫评；接受为低成本审查候选，本轮未实现、未实验，亦不自动进入执行/evolve。" }, { "id": "2610.20413", "title": "When Does Momentum Help in Reinforcement Learning with Verifiable Rewards?", "batch": "2026-10-15-review-closure-post-training", "resolution": "affordable_candidate_not_implemented", "reason": "只接收论文原始十臂 bandit 的 bias-corrected EMA/安全 η/(1−β) 机制和解析性质检验，可用 CPU 数值运算保留原目标；不接收4B/32B RLVR或50,000-replicate完整统计活动，不据此声称LLM提升；接受为低成本审查候选，本轮未实现、未实验，亦不自动进入执行/evolve。" } ], "accepted_catalogued": [ { "id": "2610.11167", "key": "pivot", "detail_path": "post-training/2610.11167-pivot/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11247", "key": "opd-learning-signals", "detail_path": "post-training/2610.11247-opd-learning-signals/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11251", "key": "vcola", "detail_path": "multimodal-models/2610.11251-vcola/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11291", "key": "semi-opd", "detail_path": "post-training/2610.11291-semi-opd/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11332", "key": "recal", "detail_path": "post-training/2610.11332-recal/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11519", "key": "residual-advantage", "detail_path": "post-training/2610.11519-residual-advantage/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11659", "key": "dial-opd", "detail_path": "post-training/2610.11659-dial-opd/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.11854", "key": "grpo-dropout", "detail_path": "post-training/2610.11854-grpo-dropout/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.12061", "key": "race", "detail_path": "agent-research/2610.12061-race/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.12367", "key": "sguid", "detail_path": "post-training/2610.12367-sguid/README.md", "document_exists": true, "adapter_contract_type": "str", "capability_completeness_verified": false }, { "id": "2610.12417", "key": "woven", "detail_path": "reproductions/2610.12417-woven/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.12707", "key": "umrm", "detail_path": "reproductions/2610.12707-umrm/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13025", "key": "evo2", "detail_path": "reproductions/2610.13025-evo2/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13092", "key": "jev-workflow", "detail_path": "reproductions/2610.13092-jev-workflow/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13194", "key": "janus", "detail_path": "reproductions/2610.13194-janus/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13257", "key": "cast-opd", "detail_path": "reproductions/2610.13257-cast-opd/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13285", "key": "boundary-evo", "detail_path": "reproductions/2610.13285-boundary-evo/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13348", "key": "t2mem", "detail_path": "reproductions/2610.13348-t2mem/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13481", "key": "branch-opd", "detail_path": "reproductions/2610.13481-branch-opd/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13521", "key": "road-mopd", "detail_path": "reproductions/2610.13521-road-mopd/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13630", "key": "msar", "detail_path": "reproductions/2610.13630-msar/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13735", "key": "orthros", "detail_path": "reproductions/2610.13735-orthros/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13767", "key": "dualquant", "detail_path": "reproductions/2610.13767-dualquant/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13806", "key": "forte", "detail_path": "reproductions/2610.13806-forte/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13889", "key": "rated", "detail_path": "reproductions/2610.13889-rated/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false }, { "id": "2610.13912", "key": "rsi-anything", "detail_path": "reproductions/2610.13912-rsi-anything/README.md", "document_exists": true, "adapter_contract_type": "dict", "capability_completeness_verified": false } ], "accepted_closed": [ { "id": "2610.02626", "title": "Imagine the Future, Internalize the Gist: Efficient VLA Reasoning via Internalized Spatiotemporal Imagination", "terminal_observations": [ { "batch": "2026-10-15-known-revisions", "resolution": "not_reproduced_high_cost", "reason": "IG-VLA v3先学习未来SigLIP状态，再将时空reasoning内化为gist；不是仅重复当前帧。§4.1原文训练reasoning用8 H100、gist用4 A6000，LIBERO 100K步和VLABench61K步。成本高，不复现、审查 DONE，单卡推理延迟结果不替代训练验收。" } ] }, { "id": "2610.05417", "title": "Render to Reason: Novel-View Semantic Prediction Improves Spatial Understanding in VLMs", "terminal_observations": [ { "batch": "20261020-harness2-priority-historical-repair", "resolution": "not_reproduced_high_cost", "reason": "成本过高，不复现（审查 DONE）。AppendixA.1:8GH200;4B variants12h,InternVL8B30h,Qwen3VL8B96h per training configuration, plus ablations." } ] }, { "id": "2610.07338", "title": "Logbook: Extremely Long-form Audio Event Understanding", "terminal_observations": [ { "batch": "20261020-harness2-priority-historical-repair", "resolution": "not_reproduced_high_cost", "reason": "成本过高，不复现（审查 DONE）。52 systems; Table1 test audio82.1h+240.6h+148.9h, full cascaded/LLM fact+entailment evaluation. Full campaign high cost." } ] }, { "id": "2610.08077", "title": "Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight", "terminal_observations": [ { "batch": "2026-10-15-known-revisions", "resolution": "not_reproduced_high_cost", "reason": "SRD v3用事后经验监督行动前不可见轨迹的foresight，推理无需生成foresight；训练时可用hindsight不等于评测agent可读未来。Appendix C.2为4B/9B thinking-on、8 H200、每步128轨迹及十种agent任务，成本高，不复现、审查 DONE。" } ] } ], "legacy_deferred_reason_categories": { "explicit_unimplemented": 6, "implementation_or_execution_preconditions": 20, "integrity_or_acceptance_audit_preconditions": 3, "reference_only": 1, "further_evidence_verification": 1 }, "legacy_deferred_not_structured_accepted": [ { "id": "2610.09027", "title": "Visual Memory Attacks Can Persist Through The KV Cache", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "P-VMI whitebox图像攻击移除后仍经后续KV影响，cache-swap/summarycompaction控制；真KV交互/梯度攻击待实现，onlysoftprompt是diagnostic。" }, { "id": "2610.09127", "title": "CADFather: Autonomous CAD Reconstruction through Coordinated Tool Use", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读CADFather method/experiment：VLM看target/intermediate renders，协调pretrainedCADgeneration/numericalrefinement，真CadQueryDSLexecute feedback与bestvalidarchive，无新training。下一批须真实3D几何/多tool策略，注意testsubsetablation不能参与parameter选择。" }, { "id": "2610.09159", "title": "SpecGuard: Proving a Task Is Broken Before the Agent Cheats", "frozen_review_decision": "reviewed_deferred", "reason_category": "integrity_or_acceptance_audit_preconditions", "reason": "Google DeepMind优先。已读SpecGuard方法、实验和限制：SpecAgent独立测试生成Lean specification，TestAgent分离，semantic connector与kernel证书只有在忠实autoformalization下保证。349 mutated SWE任务和手工60规范审计，不是线上A/B；下一批须防规范弱化/伪冲突。" }, { "id": "2610.09252", "title": "Adaptive Visual Token Reduction for Accelerated Image Understanding", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "ReFIT multi-scale instructionwindowreshape+tokenrefine，只训reductionmodules；四VQA FLOPs非wallclock提升，真实feature模块待实现。" }, { "id": "2610.09430", "title": "SkillCycle: Co-Evolving Agent Policies and Skill Banks", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "SkillCycle bank/router冻结policy学习后冻结policy修rules，rule/bank环境配对验证；ALFWorld/WebShop真实controller/distill待实现，no-skill部署不可带bank。" }, { "id": "2610.09484", "title": "MIMESIS: Learning User Simulators as Training Environments for Interactive Agents", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "Meta优先。已读MIMESIS user simulator及agent posttraining：user-side midtrain、ThoughtTrace SFT、多域GRPO后冻结simulator，再agent GRPO+CSD token重加权；private thought仅retrospective teacher训练上下文。9 unseen user评测和8环境，下一批须真实双训练链而非fixed simulator。" }, { "id": "2610.09624", "title": "How Do Agentic LLMs Decide to Call Tools? A Tool-Call Vector Shaped by Suppression", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读toolcall causalvector分析/跨model评测：500minimalverbpairs300discover200heldout，hiddenstateintervention/transcoder，7families与native tau2 generalization。下一批须真实activations causalnecessity/sufficiency，不用文字heuristic决定toolcall或按test筛vector。" }, { "id": "2610.09761", "title": "PARC-Loc: Text-to-Point-Cloud Localization with Partial Assignment and Relational Consistency", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "PARCpartialassignment+pairwise spatial关系和neighborboundary扩充，KITTI360/CityLoc；真实text/pointcloud特征与coarse/fine权重待实现。" }, { "id": "2610.09856", "title": "Self-Evolve With a Reference:Anchored Training of Tool-Integrated Agents", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读AnchorLoop方法/实验：历史Executor frozenanchor同时给cross-referenceadvantage和Curriculum一致性proxy，VeRL sandbox+交替真实训练；curriculum时anchor与executor同参数，不能冒称版本改善或correctness。下一批须保留dualagent训练和reference snapshot，不以selfconsistency当gold。" }, { "id": "2610.09935", "title": "AgentTracer: Tracing Indirect Prompt Injection Attack through Fine-Grained Intention-Execution Alignment", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读AgentTracer graphdesign/evaluation：userintent authorization与operationKB、taskintent implicitdependency、从anomalouscall backwardprune追injectionsource；successattacklogs加背景noise。下一批须完整trace与无attackfalsepositive控制，不能已知injection位置当attributor输入。" }, { "id": "2610.09952", "title": "MSU Team at the Explainable Deepfake Detection Challenge 2026: Grounded Artifact Evidence for Deepfake Detection", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "XPlainVerse DINO/Mesorchdetector、trainingartifactmask和patchcontrastive+QwenVL解释/GRPO简化；publictrainval/test可验，真多stage待实现。" }, { "id": "2610.10062", "title": "Loud Failures, Quiet Failures: Fault Detection and Recovery in Tool-Using Language Model Agents", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读loud/quietfailure method/results：自有BFCL statefultool layer四typedfault控制注入，24task1920trial、faultfree nondeterminism基线与检测/改plan/recovery分开。下一批须真实tool状态execute，gold finalstate仅scorer，reasoning无recovery提升负结果保留。" }, { "id": "2610.10071", "title": "HGP:An on-device personalized agent memory via hybrid graph storage", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读HGP方法/实验：confidence-driven LLMannotation在线训lightweightrouter、episode/semantic/proceduralgraphs+workingstate；LoCoMo/PAL选择与QA。下一批须真实router更新与图检索、on-device latency/资源实际测，不能只标签分类fixture。" }, { "id": "2610.10120", "title": "RewardWeaver: Long-Horizon Interactive Learning for Language Agents via Self-Evolving Reward Adaptation", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读RewardWeaver method/experiment：admitted rubric语义固定，阶段训练后lowoutcome backwardattribution动态reward allocation，独立controlled expansion；社会/议价/sales。下一批须实际policy训练、validatedrubric稳定，不能追分随意重定义reward或测试goal泄漏。" }, { "id": "2610.10305", "title": "How Do Transformers Learn to Represent Symmetries?", "frozen_review_decision": "reviewed_deferred", "reason_category": "explicit_unimplemented", "reason": "Transformerfiniteaugmentation学习pointcloud群不变性与equivariance，受控architecture/ablation；真实训练/OOD几何证据非fixture待实现。" }, { "id": "2610.10786", "title": "Plan-and-Patch: Diffusion Language Models for Agentic Planning", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读Plan-and-Patch方法/实验：真实diffusion planner parallelunmask、region infill保prefix/suffix，与独立executor失败region选择，NaturalPlan/ALFWorld/TextCraft训练。下一批须dLLM与联合generation/repair训练，AR字符串编辑非定义机制。" }, { "id": "2610.11008", "title": "Why LLM Agents Favor Their Group: Stakes, Observed Norms, and Reputation", "frozen_review_decision": "reviewed_deferred", "reason_category": "reference_only", "reason": "已读群体favoritism实验/结果：任意group、costless/stake、scripted history与个体reputation控制，4400模拟societies。属于agent社会行为测量，保留下一批参考；不能把合成群体pointallocation解释为现实组织因果。" }, { "id": "2610.11123", "title": "When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读GenUI task定义/实验：ToolAgent与GUICoder、DynamicUX实际sandbox交互reward、RewardAuditor meta-rubric，UI-TAU真实publicdatabase模拟任务与4B训练。下一批须真实UI/code执行和用户输入，judge reward hacking校验；rounds问卷不当生产A/B。" }, { "id": "2610.11136", "title": "The \"10th Juror\": Open-Set Standpoint Screening for Bureaucratic Bias Detection", "frozen_review_decision": "reviewed_deferred", "reason_category": "integrity_or_acceptance_audit_preconditions", "reason": "已读MARS-Gov方法/实验：legal retrieval、open-set scout动态juror、conservative routing和rewrite语义verification，DGDB/LOCO评测。下一批须独立evidence与重写含义audit，法律背景检索不可视为法律建议/无偏性保证。" }, { "id": "2610.11226", "title": "When Lower Reconstruction Loss Hurts: Distributionally Robust Refinement for Low-Bit LLM Quantization", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "DRQ固定quant grid只refine integer codes，worst-case activation distributions纠正calibration reconstruction泛化，六PTQ methods dense/MoE公开ppl/tasks；下一批须真实整数码/量化forward/uncertainty inner problem。" }, { "id": "2610.11384", "title": "Environmental Feedback Modeling Matters: Rethinking Feedback Treatment in Agentic Hindsight Self-Distillation", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读SELF方法/实验：on-policy真实environment反馈建feedback-conditioned hindsight teacher，joint response modeling loss+selfdistillation；tau/AppWorld reward-free agent训练。下一批须真实双objective梯度，privileged未来feedback仅训练teacher不进部署policy。" }, { "id": "2610.11573", "title": "Memory Type Varies: Empowering LLM Agents for Long-Term Memory with Diverse Strategies", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读MemoType/TriMEM方法和评测：learned router识别memory/query类型选择retrieval，Contriever统一backbone、LongMemEval/LoCoMo/PerLTQA。下一批须实际router训练与检索标签隔离，人工分类型heuristic不能称核心复现。" }, { "id": "2610.11578", "title": "Chronos Enables Code Agents to Reason over Software Evolution", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读Chronos方法/设置：历史PR蒸馏experience cards+typed graph、semantic entry+multi-hop、双patch生成steward选择；6backboneSWE对比。下一批须截止目标issue前history防答案PR泄漏，真实patch执行而非检索静态gold。" }, { "id": "2610.11862", "title": "LEVER: Adaptive Cost-Aware Proof Search Over AND/OR Graphs", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读LEVER AND/OR graph与实验：partial proof realized objective加open subgoal预测，programmable cost/length/impurity，Lean kernel保correct；matched model/version/price。下一批须真实Lean搜索、预测前评分与cost含失败，不能gold proof排序代替。" }, { "id": "2610.12248", "title": "EgoVoice: Proactive Spoken Assistance from Egocentric Multimodal Streams", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "EgoVoice用HoloAssist source-separated/resynth audio和listen-or-speak SFT/DPO训练主动语音助手，170 sessions onset/content/end评测；多模态/语音后训练下一批须实际流式音视频与timing policy。" }, { "id": "2610.12611", "title": "SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读SoftwareGym方法/实验：合成真实可执行GUI app、固定v1 GUI测试迭代repair，182app12类120task，Qwen3.5-9B multimodal SFT、out-of-domain OSWorld。下一批须VM实际GUI交互与合成环境自测隔离，不能用静态截图/固定动作fixture。" }, { "id": "2610.12655", "title": "CLEAT: Learning Value-Guided Clarification for Interactive Agents", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读CLEAT摘要、intent alignment及实验：同history ask/proceed paired rollout估residual query value，controller与task policy co-train，新policy outcome更新labels；Qwen3-4B四bench十设置三seed。下一批须训练真实value predictor+policy，不能拿测试latent goal作agent输入。" }, { "id": "2610.12794", "title": "Meta-Reflective Evolution Effectively Improves Production AI Systems", "frozen_review_decision": "reviewed_deferred", "reason_category": "further_evidence_verification", "reason": "已读CoCoEvolve方法与实验协议：archive parent、history检索、live target probing、自测mutation再external评分；train实例可见，validation仅score、test隔离。production AI目标及公开heldout基准不自动证明生产流量A/B，下一批须进一步核验live authority和真实evaluator，非registry标签。" }, { "id": "2610.12957", "title": "JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读JoyInside方法/实验：产品条件multimodal trajectory、SFT、多teacher OPD及DPER token-level privileged evidence路由；部署单student不读teacher reference。私有120k混合数据、领域skill及两机器人case限制公开复现，下一批须保留真实多阶段训练。" }, { "id": "2610.13144", "title": "Competence-Aware Orchestration for LLM-Based Multi-Agent Systems", "frozen_review_decision": "reviewed_deferred", "reason_category": "implementation_or_execution_preconditions", "reason": "已读competence orchestration方法/实验：固定probe capability fingerprint加query-conditioned模型，learned role activation与communication graph；6 reasoning/codebench及unseen backbones。下一批须训练真实competence predictor，probe/test隔离而非gold推断能力。" }, { "id": "2610.13166", "title": "Scaling Agentic Data for Long-Horizon Terminal Intelligence", "frozen_review_decision": "reviewed_deferred", "reason_category": "integrity_or_acceptance_audit_preconditions", "reason": "已读TerminalHorizon方法/实验：独立execution发现合成环境预先完成的关键工作，再reconstruct并扩展同goal任务；1.5k环境、Qwen3.5-27B SFT与12bench。production settings是任务评测不证明线上A/B；下一批须隔离environment gold verifier与agent可见包。" } ], "global_title_duplicate_ids": [ { "normalized_title": "motif", "ids": [ "2608.25381", "2608.25392" ] } ], "caveats": [ "504 is explicit accepted/no catalog/no explicit terminal within the three frozen accepted registers, not all possible repository work.", "31 are reviewed_deferred research leads, not automatic implementation commitments.", "No current 521 identity duplicates, title aliases to manifest, or current delivery-path matches were found.", "Catalog status is distinct from faithful algorithm, runtime validation, Evolve integration and measured capability.", "priority_organizations returns empty for all old pending rows; reason-only organization signals are not recognized." ] }