提出可解释的内存保留机制，让大模型在长任务中自动筛选有用记忆。
Selective Memory Retention for Long-Horizon LLM Agents
- 用可解释特征评分内存条目，按优先级淘汰低分项。
- 在噪声数据中保持16.6%的精准率，任务成功率97/100。
- 适合长周期任务中抗干扰、需高效记忆管理的场景。
何时记忆保留对增强型大模型代理至关重要？我们通过TraceRetain框架，在冻结的LLM代理中实现有界外部记忆，基于可解释特征（成功性、年龄、访问频率、冗余度、特异性、相似性、下游效用）为记忆条目打分，并在容量满时淘汰得分最低者。在干净的ALFWorld环境中，使用gpt-5-mini，外部记忆相比无记忆显著提升，但不同有界保留策略间差异在威尔逊95%置信区间内：在T=100至T=200期间，清洁环境未自然出现记忆污染问题。在受控噪声写入压力下（75%合成干扰项），无界记忆与FIFO-K50的Precision@5从20.2%降至12.4%和15.8%降至3.8%，而TraceRetain-CEM基本不变（16.9%→16.6%），并维持97/100的任务成功率。机制分析显示，无界记忆平均相似度最高（0.87）但精准率最低，表明失败干扰项在嵌入空间中接近查询。离线分布内评估显示，带记忆策略解决47至49/50任务，而无记忆仅39/50。有界保留策略在饱和清洁基准上提升内存与步数效率，且仅在输入流含噪声时优于传统缓存启发式方法。
原文摘要 · Abstract (English)
When does retention matter for memory-augmented LLM agents? We study this with TraceRetain, a lightweight framework for bounded external memory in frozen LLM agents that scores entries by interpretable features (success, age, access frequency, redundancy, specificity, similarity, downstream utility) and evicts the lowest-scoring ones at capacity. On clean ALFWorld with gpt-5-mini, external memory robustly improves over no memory across two seeds, but differences among bounded retention policies fall within Wilson 95% CIs: clean ALFWorld at T=100 to T=200 does not naturally exhibit the memory pollution retention is designed to address. Under a controlled noisy-write stress (75% synthetic distractors), unbounded memory and FIFO-K50 degrade on Precision@5 (20.2% to 12.4% and 15.8% to 3.8%) while TraceRetain-CEM is essentially unchanged (16.9% to 16.6%) and preserves 97/100 task success. The mechanism: unbounded memory has the highest mean similarity (0.87) but lowest precision, indicating failed distractors close to the query in embedding space. Held-out in-distribution evaluation shows memory-augmented policies solving 47 to 49 of 50 tasks vs. 39/50 for no memory. Bounded retention buys memory and step efficiency on saturated clean benchmarks at no task-success cost, and only differentiates from cache heuristics when streams contain noise.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。