Tag: interpretability

All the articles with the tag "interpretability".

I tried to catch a language model making a math mistake. My first accepted paper is about why it did not work.
Published:Oct 4, 2026 at 03:30 PM
My MATH-AI workshop paper at NeurIPS 2026 tests whether drift in a language model's hidden state can flag and fix math errors mid-solution. It does neither, and getting to that answer taught me more than a positive result would have.

I tried to catch a language model making a math mistake. My first accepted paper is about why it did not work.