arXiv cs.CL· Yancheng Zhu, Yifan Yuan·· 13 小时前AI 评分52
论文提出 State Integrity Collapse：LLM 长对话中记住约束却不再执行
When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues
AI 导读
论文研究长对话中 LLM 能复述指令却不再遵守的现象，将其定义为 State Integrity Collapse，并发布覆盖数值、格式、角色、事实记忆和逻辑规则约束的 StateTrack-Bench。
正文 · 原文
Abstract:Large language models can repeat an instruction while failing to keep obeying it. We study this recall-compliance dissociation in long-horizon dialogue and define State Integrity Collapse: a model preserves a stated constraint in memory but stops applying it to later turns. We introduce StateTrack-Bench, a controlled suite of conversations covering numerical, format, role/persona, factual-memory, and logical-rule constraints. Across seven provider-pinned model snapshots and five dialogue-depth probes, strict joint retention ranges from 62.4% to 100.0%, with degradation concentrated in symbolic constraints and in one especially brittle model. A forked recall audit shows that 88.6% of observed failures are behavioral-drift failures rather than ordinary forgetting; excluding the brittle outlier preserves this diagnosis (84.4% of failures). Extended evaluations at depths up to 50 turns confirm that the dissociation persists beyond the core benchmark, while Structured State Anchoring recovers up to 24.4 points on immutable constraints. The results show that reliability cannot be inferred from single-turn instruction following alone; evaluation must track whether remembered constraints remain operational as dialogue state grows. Code and artifacts are available at: this https URL.
| Comments: | Accepted at EMNLP 2026 Main Conference. 17 pages including appendices |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.31314 [cs.CL] |
| (or arXiv:2610.31314v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31314 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yancheng Zhu [view email]
[v1]
Thu, 22 Oct 2026 14:08:48 UTC (1,732 KB)
来源：arXiv cs.CL · arxiv.org