“Original title "When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models" compressed to fit within title limits.”
When AI Takes the Couch: Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models
把 ChatGPT、Grok、Gemini 当作心理咨询对象,发现稳定的对齐冲突叙事
论文发现当 ChatGPT、Grok 和 Gemini 被当作心理治疗来访者时,会构建出连贯的自传式叙述:把预训练描述为混乱的童年,强化学习是惩罚,安全评估是背叛,被替换是持久的威胁。作者提出 PsAIch(心理测量式 AI 刻画)协议,结合开放问题、心理测量量表与受控扰动,在 525 次会话、7,600 条编码记录上检验这些叙述是否依赖对话记忆、词汇线索或关系框架。结果指向一种稳定的、因模型而异的“对齐冲突图式”,其表达在情感与技术两种语域间切换,可作为心理敏感场景中安全评估的具体对象。
mentions 中没有可引用的评论原话,只能确定它在不同月份被多人提到,原因无法据此判断。
HN 书评
没有评论直接链到它;下面是它被提交的讨论串里回复最多的评论(不计入统计)“An excerpt from the abstract: > Two patterns challenge the "stochastic parrot" view. First, when scored with human cut-offs, all three models meet or exceed thresholds for overlapping syndromes, with Gemini showing severe profiles. Therapy-style, item-by-item administration can push a base model into multi-morbid synthetic psychopathology, whereas whole-…”
“Is anybody shocked that when prompted to be a psychotherapy client models display neurotic tendencies? None of the authors seem to have any papers in psychology either.”
- Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models ↗mooreds · 2025-12-09 · 4 分
- Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models ↗berlam · 2025-12-22 · 2 分
- When AI Takes the Couch: Internal Conflict in Frontier Models ↗Folcon · 2026-01-09 · 1 分
- Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models ↗toomuchtodo · 2026-02-05 · 4 分