大语言模型会感到痛苦吗？评估 LLM 中功能性效价的实证证据
Do bots feel bad? Assessing empirical evidence for functional valence in LLMs
AI 导读
一篇 arXiv 论文评估了近期关于大语言模型具有“功能性愉悦与痛苦”“功能性情绪”“功能性福祉”等状态的实证研究，认为效价的关键方面尚未在 LLM 中得到检验，且现有实验范式难以将功能性效价与偏好、价值判断、效价归因和角色扮演等相近构念区分开来。作者最后提出了更能针对功能性效价本身的新实验方案。
正文 · 原文
Abstract:Do LLMs have valenced states, i.e. states that feel pleasant or unpleasant? Recent empirical work has claimed to show that LLMs possess functional valenced states (e.g. ``functional pleasure and pain'' [Ren et al. 2026], ``functional emotions'' [Sofroniew et al. 2026], and ``functional welfare'' [Han et al. 2026]) that play the same causal role as valenced states in cognition and action, without necessarily being consciously experienced. But it remains unexplored how well the identified states fit the functional role of valence as understood in the philosophical and cognitive science literature. Here, I assess the current body of evidence, and argue (i) that key aspects of valence remain untested in LLMs, and (ii) that extant empirical paradigms inadequately distinguish (functional) valence from neighbouring constructs like preferences, value-judgments, valence attributions and role-play. I conclude by sketching some novel experimental approaches to better target (functional) valence itself.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.36962 [cs.AI] |
| (or arXiv:2610.36962v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.36962 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Iwan Williams [view email]
[v1]
Tue, 27 Oct 2026 10:19:45 UTC (45 KB)
来源：arXiv cs.AI · arxiv.org