A mind-expanding conversation on how models actually work | Emmanuel Ameisen (Anthropic)
Anthropic interpretability researcher Emmanuel Ameisen joins Lenny Rachitsky to pull back the curtain on how large language models think, compute, and plan. They explore why next-token prediction produces multi-step internal representations, why anthropomorphizing AI is practically useful, how smooth scaling curves produce the illusion of sudden capability jumps, and how interpretability serves as a vital safeguard for AI alignment.
- Anthropic interpretability researcher Emmanuel Ameisen explains that AI models follow smooth, predictable scaling curves rather than erratic, sudden leaps in intelligence. Read →
- Training an AI model produces capability without explanation. As Emmanuel Ameisen puts it, "to make them is not to understand them." Read →
- Anthropic interpretability researcher Emmanuel Ameisen inspects internal activation states to catch thoughts models never output into text. Read →
- Emmanuel Ameisen rejects the phone keyboard autocomplete mental model because predicting the next token in hard tasks requires modeling underlying real-world systems. Read →
- Emmanuel Ameisen at Anthropic studies internal neural representations to see how models make decisions under the hood. Read →
- Anthropic interpretability researcher Emmanuel Ameisen argues that viewing large language models as deterministic software programs misleads engineers about how systems fail. Read →