Token-in-context feature

Creator

Creator

Created

Created

2025 Jan 30 0:40

Editor

Editor

Edited

Edited

2025 Jan 30 1:37

Refs

Refs

Pure context features or token features in dictionaries with few learned features, but then splitting into token-in-context features as more features are learned.

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning

Mechanistic interpretability seeks to understand neural networks by breaking them into components that are more easily understood than the whole. By understanding the function of each component, and how they interact, we hope to be able to reason about the behavior of the entire network. The first step in that program is to identify the correct components to analyze.

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning

https://transformer-circuits.pub/2023/monosemantic-features#phenomenology-feature-motifs

https://arxiv.org/pdf/1906.02715

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning

Mechanistic interpretability seeks to understand neural networks by breaking them into components that are more easily understood than the whole. By understanding the function of each component, and how they interact, we hope to be able to reason about the behavior of the entire network. The first step in that program is to identify the correct components to analyze.

Towards Monosemanticity: Decomposing Language Models With Dictionary Learning

https://transformer-circuits.pub/2023/monosemantic-features#discussion-token-in-context

Recommendations

////////////