AI Tools & Models HOTFeaturedAll
arXiv cs.AIScore60

All Quiet on the Neural Front: Language Models Do Not Report Injected Pain States

Summary

This study tests whether a language model can report a pain state it represents, finding that even with an amplifiable pain direction, the model does not report it.

Why this matters

Covers model · Relevant to this topic

Open original ↗

Source: arXiv cs.AI · arxiv.org

Topics

Related items