Can the Assistant Axis Predict Reward-Hacking?

·LessWrong··

NOTE: This article was written by Jansen Nye, a member of my research lab based on some exploratory studies we've done. As he notes, we collaborated on this research. The article was written by him with some edits suggested by me. -- KobyTL;DR The Assistant Axis doesn't seem to have any special connection to reward hacking, which we show. We suspected that a drop in the Assistant Axis might be a symptom of the Persona Selection Model breaking down in the face of RLVR. We also discuss pitfalls of...

Read full article →

Related Articles

Anthropic S-1
SiNTEx · Hacker News · 6h ago
Brown student maps the hidden global labor powering AI
cdrnsf · Hacker News · 6h ago
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
minimaxir · Hacker News · 22h ago
A 10-year-old CNN outperforms frontier LLMs for skin cancer diagnosis
sungam · Hacker News · 1d ago
Fixing Tool-Call Accuracy in vLLM
coffeemug · Hacker News · 6h ago