✳flâneur — a map of the web's best reading

Tracing the Thoughts of a Large Language Model — LessWrong

lesswrong.com · 6,971 words · saved by 1 readers

[This is our blog post on the papers, which can be found at https://transformer-circuits.pub/2025/attribution-graphs/biology.html and https://transformer-circuits.pub/2025/attribution-graphs/methods.html.] Language models like Claude aren't programmed directly by humans—instead, they‘re trained on large amounts of data. During that training process, they learn their own strategies to solve problems. These strategies are encoded in the billions of computations a model performs for every word it writes. They arrive inscrutable to us, the model’s developers. This means that we don’t understand how models do most of the things they do. Knowing how models like Claude think would allow us to have a better understanding of their abilities, as well as help us ensure that they’re doing what we intend them to. For example: We take inspiration from the field of neuroscience, which has long studied the messy insides of thinking organisms, and try to build a kind of AI microscope that will let us i

x Tracing the Thoughts of a Large Language Model — LessWrong Interpretability (ML & AI) AI Curated 2025 Top Fifty: 14 % 308 Tracing the Thoughts of a Large Language Model by Adam Jermyn 27th Mar 2025 AI Alignment Forum Linkpost for www.anthropic.com 13 min read 23 308 Ω 106 [This is our blog post on the papers, which can be found at https://transformer-circuits.pub/2025/attribution-graphs/biology.html and https://transformer-circuits.pub/2025/attribution-graphs/methods.html .] Language models like Claude aren't programmed directly by humans—instead, they‘re trained on large amounts of data. Du

Explore this link on the map →

related reading