arXiv cs.AI· Zhiqi Wang, Yichi Zhang, Dongwon Lee, Nathalie Baracaldo, Yuchen Yang·· 3 天前AI 评分62
arXiv 论文：上下文压缩会丢弃 LLM 系统的附带约束，CONPACT 评测与提取器提出
Lost in Context Compaction: When LLM Systems Drop Side Constraints
AI 导读
arXiv 论文《Lost in Context Compaction: When LLM Systems Drop Side Constraints》研究上下文窗口承压时压缩过程会静默丢弃用户下达的附带约束（Side Constraints，如“在我确认前不要删除任何邮件”）。
正文 · 原文
Abstract:When the context window is under pressure, LLM systems compact prior context to continue ongoing tasks. We study Side Constraints (SCs), user-issued instructions such as "do not delete any emails until I confirm," that constrain an LLM's behavior across subsequent turns but can be silently dropped during compaction. In an inbox experiment, an agent deleted emails without approval in 30/100 sessions after compaction, compared with none before compaction. To systematically evaluate SC preservation, we introduce CONPACT, an evaluation suite measuring semantic retention and downstream compliance across three scenarios: multi-turn chat, multi-turn tool use, and long-horizon research.
The evaluated compactors retain only 17% of injected SCs on average. Loss occurs across compactors and settings, with retention varying by compactor, context length, SC framing, and injection location. We propose an SC extractor that maintains constraints separately from the task summary and appends them to the compacted context, achieving over 90% retention across all three scenarios without modifying the compactor or LLM. Code and data are available at this https URL.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2608.11242 [cs.CL] |
| (or arXiv:2608.11242v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2608.11242 arXiv-issued DOI via DataCite |
Submission history
From: Zhiqi Wang [view email]
[v1]
Fri, 31 Jul 2026 16:04:00 UTC (373 KB)
[v2]
Sun, 11 Oct 2026 01:17:51 UTC (1,321 KB)
来源：arXiv cs.AI · arxiv.org