SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
Paper • 2503.09532 • Published
This repository contains the artifacts related to the paper SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability.
The code can be found at https://github.com/adamkarvonen/SAEBench. The interactive interface can be found at https://saebench.xyz.