michaelwaves

Suc­cess Per Tokens

The Cookie Mon­ster Ex­plains AI Safety

VFUSE: Viru­lent Fea­ture Un­der­stand­ing With Sparse AutoEncoders

RFDiffu­sion3: A Brief Exploration

SAEBER: Sparse Au­toen­coders for Biolog­i­cal En­tity Risk

Your Mom is a Chimera

Data-Cen­tric In­ter­pretabil­ity for LLM-based Multi-Agent Re­in­force­ment Learning

AI Mood Ring: A Win­dow Into LLM Emotions

How To De­ploy a (Tiny) AI

I gave LLMs emo­tional damage

AI Teddy Bears: A Brief Investigation

The Eco­nomics of Re­plac­ing Call Cen­ter Work­ers With AIs

Les­sons from build­ing a model or­ganism testbed

A Work­flow for Sys­tem Prompted Model Organisms

Break­ing GPT-OSS: A brief investigation

michaelwaves’s Shortform