x

AI ALIGNMENT FORUM

AF

camilablank — AI Alignment Forum

camilablank

camilablank

Message

13

2mo

camilablank

13

2mo

Towards surfacing model algorithms with meta-tokens in the J-Space

by agam_bhatia, camilablank, and Neel Nanda

TL;DR We used J-lens on Qwen3.6-27B to find “meta-tokens”: tokens that surface non-obvious computation in the model. When the model reads ambiguous text, 什么意思 ("what does this mean") fires in the J-space, and steering it away makes the model answer "a boiled egg every morning is hard to beat" with...