Circuit Tracer

Anthropic's open-source Circuit Tracer helps researchers understand LLMs by visualizing internal computations as attribution graphs. Explore on Neuronpedia or use the library. Aims for AI transparency.