From the 1 of 5 linked papers with an AI index.
5 papers
Choosing Where and How to Moderate: End-to-End Trade-offs in Filter Placement and Response Rewriting
Mengya Hu, Susie Park, Suzana Ilic +5
The paper studies how to best place and combine content‑moderation filters and response rewriting in conversational systems, measuring overall usefulness and harmful exposure rathe…
Structuring Sparsity: Block-Sparse Featurizers Capture Visual Concept Manifolds
Thomas Fel, Matthew Kowal, Mozes Jacobs +22
What is the geometry of a visual percept? The most widely used protocols for decomposing neural network representations into interpretable parts treat concepts as isolated directio…
SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
Adam Karvonen, Can Rager, Johnny Lin +12
Sparse autoencoders (SAEs) are a popular technique for interpreting language model activations, and there is extensive recent work on improving SAE effectiveness. However, most pri…
Sparse Autoencoders Do Not Find Canonical Units of Analysis
Patrick Leask, Bart Bussmann, Michael Pearce +5
A common goal of mechanistic interpretability is to decompose the activations of neural networks into features: interpretable properties of the input computed by the model. Sparse…
LLM Circuit Analyses Are Consistent Across Training and Scale
Curt Tigges, Michael Hanna, Qinan Yu +1
Most currently deployed large language models (LLMs) undergo continuous training or additional finetuning. By contrast, most research into LLMs' internal mechanisms focuses on mode…