Tag: interpretability
All the articles with the tag "interpretability".
Metacognition in RL agents vs. LLMs vs. orchestrators
Published: at 03:05 AMThree kinds of AI systems, three places a "how am I doing?" signal can live, and why today the honest answer is mostly in the orchestrator.
What would convince me a model monitors itself?
Published: at 09:00 PMMy six-part evidence bar for claiming a language model monitors itself, and the probing and patching methods I'd use to clear each part.
What metacognition means for an AI model, concretely
Published: at 02:50 AMBreaking "metacognition" into six mechanisms you can actually test in a language model, and why the model saying "I'm not sure" is the weakest evidence of all.
I tried to catch a language model making a math mistake. My first accepted paper is about why it did not work.
Published: at 03:30 PMMy 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.