mlsanityData to 14 Oct 2026

When Does Selection Replace Extraction? A Pre-Registered Test of Agent Memory with a Typed Decision Model

arXiv 2609.34227 ↗First posted 28 Sep 2026cs.AI cs.CL cs.IRcited by 1

On the map: LLM agents and tool use → Agent memory systems

Abstract (arXiv)

Does conversational memory need LLM-extracted facts, or is selecting the right raw turns enough? Published results disagree. Extraction-based systems report gains from distilled facts. Recent studies find raw history with good ranking does as well, but disagree about whether ranking matters. We ran a pre-registered study on held-out LoCoMo conversations and LongMemEval. At a tight budget on LoCoMo, raw turns selected by a single call to Jev, a typed decision model, are non-inferior to an LLM-extraction memory (one-sided 95% bound -3.0 points against a -5-point margin). Blind human grading narrows the margin but does not change the result. Raw turns cost 3,061 times less to write, and the result holds with a second answer model. Within this study, reranking's gain shrinks as the budget grows. It adds 17.4 points on LoCoMo and 9.1 on LongMemEval when three of 30 candidates are kept. At generous budgets it adds 1.5 and 1.1, and extraction systems are more accurate. This suggests why published results disagree. At matched context, Jev selects as accurately as an LLM reranker (non-inferiority bound -2.0) at a third of the latency, and more accurately than a multi-call graph traversal. Reranking lowers correct abstention. Plans, code and graded answers are released.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Agent memory systems, one of 16 directions in LLM agents and tool use. Builds on 27 of its 39 references that are in the map. First posted on arXiv 28 Sep 2026.

Builds on

27 of 39 references.

Closest ancestors · 13

Same direction, newest first

  1. 2026Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents21 Sep
  2. 2026The Retriever Should Remember: Experience-Amortized Reranking for Long-Term Agent Memory24 Aug
  3. 2026LeanMem: Simple and Efficient Long-Term Memory for LLM Agents4 Aug
  4. 2026Zero-Mem: Zero-Token Memory Operations for LLM Agents31 Jul
  5. 2026Retain or Consolidate? Budget-Dependent Operator Selection for Language Agent Memory20 Jul
  6. 2026LazyMem: Retrieve Broadly, Construct Selectively for Efficient Long-Term Agent Memory17 Jul
  7. 2026MemReranker: Reasoning-Aware Reranking for Agent Memory Retrieval7 May
  8. 2026SimpleMem: Efficient Lifelong Memory for LLM Agents5 Jan
  9. 2025A Simple Yet Strong Baseline for Long-Term Conversational Memory of LLM Agents21 Nov
  10. 2025A-MEM: Agentic Memory for LLM Agents17 Feb
  11. 2025Zep: A Temporal Knowledge Graph Architecture for Agent Memory20 Jan
  12. 2024On the Structural Memory of LLM Agents17 Dec
1 more
  1. 2024Evaluating Very Long-Term Conversational Memory of LLM Agents27 Feb

Other references · 14

  1. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,389
  2. 2024LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memorycited by 707
  3. 2025LightMem: Lightweight and Efficient Memory-Augmented Generationcited by 180
  4. 2025On Memory Construction and Retrieval for Personalized Conversational Agentscited by 105
  5. 2026EverMemOS: A Self-Organizing Memory Operating System for Structured Long-Horizon Reasoningcited by 54
  6. 2026Memora: A Harmonic Memory Representation Balancing Abstraction and Specificitycited by 11
  7. 2026Decision Hijacking: Prompt Injection Attacks on Jev's Typed Probabilistic Decisionscited by 10
  8. 2026EMBER: Efficient Memory via Budgeted Evidence Retention for Long-Horizon Agentscited by 6
  9. 2026SmartSearch: How Ranking Beats Structure for Conversational Memory Retrievalcited by 5
  10. 2026Same Ranking, Different Winner: How Scoring Targets Shape LLM Memory Benchmarkscited by 5
  11. 2026ConvMemory v2: A Recall-Preserving Top-10 Evidence Reranker for Conversational Memory Retrievalcited by 2
  12. 2026Learning User-Aware Recall: Personalized Retrieval in Long-Term Conversational Memorycited by 2
2 more
  1. 2025Fidelity Before Structure: Verbatim Chunks Beat Lossy Artifact Extraction in Long-Conversation LLM Memorycited by 1
  2. 2026Training-Free Lexical-Dense Fusion for Conversational-Memory Retrievalcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026What Does a Harness Repair? A Preregistered Study of Visibility, Baseline Adequacy and Evaluation Defectscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Reproducing LightMem: Naive RAG Is Just as Good for Memory Management ↗
  2. 2024Where is the answer? Investigating Positional Bias in Language Model Knowledge Extraction
  3. 2026Supersede: Diagnosing and Training the Memory-Update Gap in LLM Agents
  4. 2026Causal Intervention-Based Memory Selection for Long-Horizon LLM Agents
  5. 2026Can Agents Forget When Asked? A Benchmark for Memory Unlearning in LLM Agents
  6. 2026What Survives When Activation Steering Is Used as Memory
  7. 2026Lost in Conversation or Lost in Translation? Diagnosing Multi-Turn Degradation in RAG
  8. 2026Locomo-Plus: Beyond-Factual Cognitive Memory Evaluation Framework for LLM Agents