Mnemon: Raw Records, Fast Judgments, Slow Thoughts

简体中文

L1 raw-record retrieval and bounded views only; no Jev checkpoint or local LoCoMo score.

Paper information

Field	Details
Paper	Mnemon: Raw Records, Fast Judgments, Slow Thoughts
Institution	The paper does not list an affiliation for the first author.
First public release	2026-09-28 (arXiv v1)
Author code	No public author repository found as of 2026-09-30.
Local adapter or method key	mnemon
Local reproduction code	src/auto_research/agent_research/latest_20260930_closure.py

Paper method and evidence

Retain dated raw records. A slower planner proposes searches, a fast Jev decides which records matter, and a compact view under a fixed budget feeds the original answering model.

flowchart LR
  M[Dated raw records] --> S[Planner proposes searches]
  S --> J[Fast Jev selects relevant records]
  J --> V[Bounded view for answering model]

Original paper figure

Figure 2 illustrates the original workflow. From the original paper; copyright remains with the authors.

Core equation

The reference preserves record selection and gating; tests cover type boundaries, determinism, and isolation.

Reported results

The paper's benchmarks, speed and learning curves are not measured by the local mini-suite.

Local reproduction

The three-seed reference tests check the operator, shapes, boundaries, and state invariants under a fixed fixture budget. These are diagnostic_only=true, not paper-scale outcomes or online gains.