The Sleeping Agent: What Gist-Based Context Compression Loses and Why

• URL: http://arxiv.org/abs/2608.11775v1
• Date: Wed, 12 Aug 2026 08:19:15 GMT
• Title: The Sleeping Agent: What Gist-Based Context Compression Loses and Why
• Abstract summary: We use a biologically-inspired compression framework motivated by sleep-based memory consolidation to study when compression helps and when it hurts.<n>We find gist compression substantially outperforms truncation on multi-hop reasoning and single-hop factual questions.<n>We trace this failure to a specific mechanism: the gist abstraction prompt preserves relational and event structure while discarding dates and times.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Gist-based context compression---summarising older conversation history into compact representations---is a common approach in long-horizon language model agents, yet its effect on different types of memory retrieval is poorly understood. We use Salience-Weighted Consolidation (SWC), a biologically-inspired compression framework motivated by sleep-based memory consolidation, as a diagnostic probe to study when gist compression helps and when it hurts. SWC scores conversation history by salience, partitions it into priority tiers, and applies structured gist abstraction to mid-priority content. Evaluating four conditions on all ten LoCoMo conversations---1,935 matched text-only questions in total, 1,501 used in the primary aggregate after excluding Category 5 (adversarial) questions---at temperature 0, we find a consistent task-type interaction: gist compression substantially outperforms truncation on multi-hop reasoning and single-hop factual questions, but temporal questions remain substantially harder under compression, with compressed conditions scoring well below the full-context reference on the conversations where both are evaluated. We trace this failure to a specific mechanism: the gist abstraction prompt preserves relational and event structure while discarding dates and times. A preservation analysis across all ten conversations confirms the mechanism: an approximately 20-fold increase in temporal expression preservation (3.05% to 62.39%) with a one-sentence prompt modification, while named entity and event preservation rates barely change (x1.02 and x1.11), demonstrating that the fix is a precision instrument. The prompt modification recovers +0.314 [0.254, 0.375] judge accuracy on category-2 (temporal) questions in the matched set. Code and results: https://github.com/kyrkewood/sleeping-agent.

Related papers

• Entity-Memory Graph Retrieval Improves Evidence Coverage in Long-Conversation Question Answering [0.0]
  On 1,986 questions from ten LoCoMo conversations, graph retrieval raises official evidence recall at top-k 25 from 79.7468% to 84.4842%.<n>Four paper-eligible requested configurations support empirical robustness across the tested GPT-3.5 and DeepSeek extractors on both outcomes.
  arXiv  Detail & Related papers  (2026-08-28T05:03:33Z)
• What Does Context Compression Cost an Agent? Interaction Costs Unrevealed by Task-Completion Metrics [0.913755431537592]
  We introduce a controlled runtime measurement protocol for represerving cost in a tool-using agent.<n>We vary compression severity, compare a dropping operator with a fact-acquisition operator, restore dropped state through controlled oracle interventions, and decompose tool calls into retrieval and execution.<n>Overall, compression can impose hidden interaction costs when execution-relevant state becomes absent and must be reacquired.
  arXiv  Detail & Related papers  (2026-08-17T10:21:36Z)
• Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression [18.435649248662944]
  We show that when retained text contains an answer but deleted text defines the entity needed to interpret it, we call the result referential dangling.<n>We train a compact to rank omitted sentences by whether they are needed to interpret retained text and reinsert the top-ranked candidates without support annotations at inference.
  arXiv  Detail & Related papers  (2026-08-05T07:58:53Z)
• H+ Embedding: Harmonizing Global and Token-Level Retrieval with Context-Dependent Phrases [20.2337258233724]
  H+ Embedding predicts variable-length phrase partitions, preserves uncovered tokens as singletons, and applies importance-guided unit selection with weighted MaxSim interaction.<n>Across 16 scientific, medical, and bilingual tasks, its phrase retrieval branch exceeds the global retrieval branch by 6.91 macro nDCG@10.<n>It also nearly matches Token while using 13.7% fewer document vectors and outperforms content-independent grouping rules under moderate vector budgets.
  arXiv  Detail & Related papers  (2026-07-29T01:37:45Z)
• MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations [50.24315431387575]
  We introduce MemOps, a benchmark that reformulates conversational memory as a sequence of lifecycle operations.<n>MemOps represents each memory event with a structured trace specifying its trigger, target, scope, state transition, and supporting evidence.<n>Across long-context, retrieval-based, parametric and managed-memory systems, MemOps disentangles failure modes that final-answer accuracy alone conceals.
  arXiv  Detail & Related papers  (2026-07-14T15:33:44Z)
• Context-Aware Distillation and Ablation for Text2DSL [78.29352305480285]
  We replace prompt-only synthetic generation with context-aware distillation.<n>A teacher large language model operates under an explicitly defined structured context.<n>This scales the verified PolkitBench corpus from 4,204 to 10,073 pairs.
  arXiv  Detail & Related papers  (2026-06-21T16:27:24Z)
• Query-Adaptive Semantic Chunking for Retrieval-Augmented Generation: A Dynamic Strategy with Contextual Window Expansion [0.0]
  We present Query-Adaptive Semantic Chunking (QASC), which dynamically constructs chunks by integrating queries into segmentation through three mechanisms.<n>QASC achieves an F1-score of 0.85, a relative improvement of 18-27% over fixed chunking and 8-12% over semantic and agentic alternatives.
  arXiv  Detail & Related papers  (2026-04-29T06:40:53Z)
• PERMA: Benchmarking Personalized Memory Agents via Event-Driven Preference and Realistic Task Environments [72.02445514666428]
  We introduce PERMA, a benchmark designed to evaluate persona consistency over time beyond static preference recall.<n>PerMA consists of temporally ordered interaction events spanning multiple sessions and domains, with preference-related queries inserted over time.<n> Experiments demonstrate that by linking related interactions, advanced memory systems can extract more precise preferences and reduce token consumption.
  arXiv  Detail & Related papers  (2026-03-24T14:04:11Z)
• Learning to Forget: Sleep-Inspired Memory Consolidation for Resolving Proactive Interference in Large Language Models [0.11080037957254413]
  Large language models (LLMs) suffer from proactive interference (PI): outdated information in the context window disrupts retrieval of current values.<n>We propose SleepGate, a framework that augments transformer-based LLMs with a learned sleep cycle over the key-value ( KV) cache.<n>SleepGate achieves 99.5% retrieval accuracy at PI depth 5 and 97.0% at depth 10, while all five baselines -- full KV cache, sliding window, H2O, StreamingLLM, and decay-only ablation -- remain below 18%.
  arXiv  Detail & Related papers  (2026-03-15T17:54:43Z)
• Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents [80.33280979339123]
  We introduce Memory-T1, a framework that learns a time-aware memory selection policy using reinforcement learning (RL)<n>On the Time-Dialog benchmark, Memory-T1 boosts a 7B model to an overall score of 67.0%, establishing a new state-of-the-art performance for open-source models.
  arXiv  Detail & Related papers  (2025-12-23T06:37:29Z)
• ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge [50.93758649363798]
  Impliret is a benchmark that shifts the reasoning challenge to document-side processing.<n>We evaluate a range of sparse and dense retrievers, all of which struggle in this setting.
  arXiv  Detail & Related papers  (2025-06-17T11:08:29Z)
• On Memory Construction and Retrieval for Personalized Conversational Agents [69.46887405020186]
  We propose SeCom, a method that constructs the memory bank at segment level by introducing a conversation segmentation model.<n> Experimental results show that SeCom exhibits a significant performance advantage over baselines on long-term conversation benchmarks LOCOMO and Long-MT-Bench+.
  arXiv  Detail & Related papers  (2025-02-08T14:28:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.