Chris Olah

Co-founder of Anthropic's interpretability team and, before that, a driving figure behind the Circuits thread and Distill.pub — the research tradition of treating neural-network internals as objects worth reverse-engineering that entity-on-the-biology-of-a-large-language-model descends from directly. Co-author on that paper; named individually in the vault for the first time in this session's capture, though his methodological lineage has been present implicitly since the vault's earliest note on the paper.

References

• claim-biology-llm-refusal-chain-is-harmful-request-recognition · claim-biology-llm-jailbreak-sentence-boundary-delays-not-triggers-refusal
• entity-on-the-biology-of-a-large-language-model · entity-jack-lindsey · entity-attribution-graphs