Qi Zhang

On the Limits of Sparse Autoencoders: A Theoretical Framework and Reweighted Remedy

SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning

SAE as a Crystal Ball: Interpretable Features Predict Cross-domain Transferability of LLMs without Training

Understanding and Leveraging Expert-level Monosemanticity in Finetuning MoE LLMs

Projection Head is Secretly an Information Bottleneck

Beyond Interpretability: The Gains of Feature Monosemanticity on Model Robustness