LLM Feature Representation

5 posts · 1 source

AnthropicBlog5 Oct

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning

AnthropicBlog4 May

Distributed Representations: Composition & Superposition

AnthropicBlog14 Sept

Toy Models of Superposition

AnthropicBlog17 Jun

Softmax Linear Units

AnthropicBlog8 Mar

In-context Learning and Induction Heads