Refuting behaviorist Persona: Does RL change personality via circuits?
voooooogel · x · 2026-08-31
Argues against defining persona solely by responses, noting the model has metacognitive access. Cites examples like inserting 'meow' tokens triggering ICL roleplay without touching core persona machinery (e.g., assistant axis). Questions if RL induces persona or just modifies late-layer circuits (e.g., inserting suffixes after refusals), affecting alignment, and suggests different persona prompts may operate via distinct mechanisms.
More from Research
- A beginner-friendly introduction to causal inference kicks off, Part 1 — Nasereliver · 2026-10-16
- FBI method trains for Pass@K with one sample per update at 12.3% of the cost — EmtiyazKhan · 2026-10-16
- Berkeley researchers sample humanoid robot policies directly from weight space with diffusion — berkeley_ai · 2026-10-16
- OneWM: unified causal training shows world and action modeling improve each other — IAAI233 · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16
- 5.8M-param Speech Event Detector adds laughs, breaths and背景 sounds to Whisper transcripts — Ayush Kumar Bar · 2026-10-16