Knowledge After Collapse:上下文与措辞如何影响语言模型事实可及性
Knowledge After Collapse: Context and Wording Shape Factual Accessibility in Language Models
AI 导读
一项 arXiv 研究提出 Knowledge Answer Divergence(KAD)评测协议,在语言模型采纳错误纠正后,区分其后续的坚持、恢复与拒答行为。
正文 · 原文
Abstract:When a language model accepts a false correction in a factual discussion, its next answer can depend on how the conversation continues. We introduce Knowledge Answer Divergence (KAD), an event-conditioned evaluation protocol that distinguishes initial false-answer adoption from subsequent persistence, recovery, and abstention. KAD freezes the conversation at the first explicit adoption of a designated false target and branches independent probes from that exact history. This enables matched comparisons of post-adoption behavior and history-management interventions that a single susceptibility score cannot capture. We evaluate eight local checkpoint conditions and six hosted endpoints, alongside an audited 300-question rerun. In corrected Llama 3.2 Instruct 3B controls, a neutral-history branch yields 42/46 correct responses, compared with 16/46 under full history. A matched cue-placement control changes false adoption by 27.3 percentage points in Qwen 3.5 4B. Twelve independently authored pressure templates demonstrate substantial sensitivity to wording. Together, these findings demonstrate the utility of KAD for evaluating factual reliability beyond initial susceptibility, enabling systematic comparisons of false-answer persistence, recovery conditions and conversational interventions.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.17763 [cs.AI] |
| (or arXiv:2611.17763v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.17763 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sujato Dutta [view email]
[v1]
Fri, 13 Nov 2026 05:19:42 UTC (1,325 KB)
来源:arXiv cs.AI · arxiv.org