Researchers debate LLM layerwise picture: where does a model's persona live?
maxsloef · x · 2026-11-01
Max Sloef endorses a layerwise picture of LLMs — early layers translate tokens into concepts, middle layers compute over them, late layers translate back to tokens — arguing there's good evidence for it. But the open question he's pursuing is where a model's persona/character is decided. Ivan Vendrov suggests testing the hypothesis with Goodfire's interpretability tools.
More from Research
- Steering just 100 of Qwen3-8B's 440k neurons matches persona vector effect — maxsloef · 2026-11-01
- Will the replication crisis slow AI progress? Burned-in intuitions may stay wrong — moultano · 2026-11-01
- Frontier LLMs still fail basic molecular cloning, researcher says — anshulkundaje · 2026-11-01
- After42: experts rate 45 of AI's 135 proposed research questions genuinely interesting — james_y_zou · 2026-11-01
- Google tests two screening policies to cut noise from agentic automated bug repair — burkov · 2026-11-01
- AI novelty checks reject 96% of ideas human scientists later publish, UIUC paper finds — rohanpaul_ai · 2026-11-01