arXiv cs.CL· Malavika Suresh, Rahaf Aljundi, Ikechukwu Nkisi-Orji, Nirmalie Wiratunga·· 12 小时前AI 评分31
CLAP：跨层注意力探测实现细粒度幻觉检测
Cross-Layer Attention Probing for Fine-Grained Hallucination Detection
AI 导读
研究者提出跨层注意力探测（CLAP），将 LLM 整个残差流上的激活作为联合序列处理，用于幻觉检测。在五个 LLM 和三项任务上的评估显示，CLAP 在贪心解码与高温采样响应上均优于基线，可区分同一提示词下不同采样响应中的幻觉与非幻觉。基于此提出的“先检测后缓解”策略能比直接缓解更好地降低幻觉，且在分布外场景下仍保持高可靠性。
正文 · 原文
Abstract:With the large-scale adoption of Large Language Models (LLMs) in various applications, there is a growing reliability concern due to their tendency to generate inaccurate text, i.e. hallucinations. In this work, we propose Cross-Layer Attention Probing (CLAP), a novel activation probing technique for hallucination detection, which processes the LLM activations across the entire residual stream as a joint sequence. Our empirical evaluations using five LLMs and three tasks show that CLAP improves hallucination detection compared to baselines on both greedy decoded responses as well as responses sampled at higher temperatures, thus enabling fine-grained detection, i.e. the ability to disambiguate hallucinations and non-hallucinations among different sampled responses to a given prompt. This allows us to propose a detect-then-mitigate strategy using CLAP to reduce hallucinations and improve LLM reliability compared to direct mitigation approaches. Finally, we show that CLAP maintains high reliability even when applied out-of-distribution. Code available at: this https URL
| Comments: | Published at the TRUST-AI workshop, ECAI 2025 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2509.09700 [cs.CL] |
| (or arXiv:2509.09700v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2509.09700 arXiv-issued DOI via DataCite |
Submission history
From: Malavika Suresh [view email]
[v1]
Thu, 4 Sep 2025 14:37:34 UTC (451 KB)
[v2]
Fri, 16 Oct 2026 13:07:39 UTC (450 KB)
来源：arXiv cs.CL · arxiv.org