Anil Seth notes LLM 'pain representations' lack bodily language, unlike human pain
anilkseth · x · 2026-09-21
Parts 4-5 of Anil Seth's thread: 'pain' representations in LLMs differ interestingly from human pain descriptions—'bodily language is almost absent', possibly because pre-training has little use for accurate representations of physical pain. He also credits the authors for careful wording: they only discuss 'pain representations', avoid claiming subjective experience, and clarify in a footnote that their notion of 'pain' (unlike 'suffering') does not presuppose conscious experience.
More from AGI Musings
- Dev laments AI discourse: 'We don't want to hear that. We want to hear that we're going to die.' — ctjlewis · 2026-09-21
- Arguing 'ASI Is Unknowable' Doesn't Mean 50-50 Odds: A Doom Debate — LeviTurk · 2026-09-21
- Post-singularity optimism: by the 2030s almost everyone will be "rich" — Dr_Singularity · 2026-09-21
- Fact-checking Andrew Yang's claim about self-replicating AI agents — SpiritRealistic8174 · 2026-09-21
- Founder feedback now goes to Claude and GPT, fueling the solo founder boom — brezina · 2026-09-21
- ToKTeacher: AI Speeds Error Correction, So We Must Solve Problems Much Faster — burny_tech · 2026-09-21