跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.10721 · v1

Behavioral Evaluation of Large Language Models Under Cognitive Biases in HAZOP Studies

首次提交 2026-11-07 21:30
版本更新 2026-11-07 21:30
雷达观测 2026-11-10 11:14
arXiv ↗PDF ↗

摘要

Hazard and Operability (HAZOP) studies safeguard industrial processes against catastrophic accidents. Large language models (LLMs) increasingly assist HAZOP drafting, yet evaluations rely on lexical overlap. Because LLMs internalize human heuristics, prompts may trigger biased judgments. Whether frontier LLMs withstand cognitive perturbations in safety-critical workflows remains unknown. Here we show across 420 trials that three frontier LLMs resisted 96.8% of cognitive manipulations, rendering overall susceptibility statistically inconclusive. Contrary to assumed broad fragility, models rejected confirmation and availability cues. However, safety-critical risk downgrades emerged under low anchors and gain frames. Thus, automated hazard analysis requires disciplined human-in-the-loop validation.

作者备注:8 pages, 1 figure, 3 tables. Submitted to IFAC for possible publication. Supplementary material available at https://github.com/ViniMelloeMuller/psyhazop-supplementary-material

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-07 21:30 · 最新观测版本

Behavioral Evaluation of Large Language Models Under Cognitive Biases in HAZOP Studies

Hazard and Operability (HAZOP) studies safeguard industrial processes against catastrophic accidents. Large language models (LLMs) increasingly assist HAZOP drafting, yet evaluations rely on lexical overlap. Because LLMs internalize human heuristics, prompts may trigger biased judgments. Whether frontier LLMs withstand cognitive perturbations in safety-critical workflows remains unknown. Here we show across 420 trials that three frontier LLMs resisted 96.8% of cognitive manipulations, rendering overall susceptibility statistically inconclusive. Contrary to assumed broad fragility, models rejected confirmation and availability cues. However, safety-critical risk downgrades emerged under low anchors and gain frames. Thus, automated hazard analysis requires disciplined human-in-the-loop validation.

查看此版本解读 →查看此版本原文 ↗