Analyze Feature Flow to Enhance Interpretation and Steering in Language Models

Daniil Laptev, Nikita Balagansky, Yaroslav Aksenov, Daniil Gavrilov

Analyze Feature Flow to Enhance Interpretation and Steering in Language Models: 59 upvotes on Hugging Face Daily Papers, #1 of 22 papers on 2025-02-07. Day-by-day upvote history.

We introduce a new approach to systematically map features discovered by sparse autoencoder across consecutive layers of large language models, extending earlier work that examined inter-layer feature links. By using a data-free cosine similarity technique, we trace how specific features persist, transform, or first appear at each stage. This method yields granular flow graphs of feature evolution, enabling fine-grained interpretability and mechanistic insights into model computations. Crucially, we demonstrate how these cross-layer feature maps facilitate direct steering of model behavior by amplifying or suppressing chosen features, achieving targeted thematic control in text generation. Together, our findings highlight the utility of a causal, cross-layer interpretability framework that not only clarifies how features develop through forward passes but also provides new means for transparent manipulation of large language models.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.