Feature-Guided SAE Steering for Refusal-Rate Control using Contrasting Prompts
"稀疏自编码器" 相关文章
What's In My Human Feedback? Learning Interpretable Descriptions of Preference Data
Time-Aware Feature Selection: Adaptive Temporal Masking for Stable Sparse Autoencoder Training
Does higher interpretability imply better utility? A Pairwise Analysis on Sparse Autoencoders
Anthropic's JumpReLU training method is really good
这次，LLM黑盒被LLM打开了~
GPT and Prejudice: A Sparse Approach to Understanding Learned Representations in Large Language Models
GPT and Prejudice: A Sparse Approach to Understanding Learned Representations in Large Language Models
Towards Interpretable and Inference-Optimal COT Reasoning with Sparse Autoencoder-Guided Generation
Towards Interpretable and Inference-Optimal COT Reasoning with Sparse Autoencoder-Guided Generation
AbsTopK: Rethinking Sparse Autoencoders For Bidirectional Features
Sparse Autoencoders Make Audio Foundation Models more Explainable
Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement
Measuring Sparse Autoencoder Feature Sensitivity
Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders
Researchers glimpse the inner workings of protein language models
Low-resourced languages get jailbroken more. Can SAEs explain why?
Rethinking Sparse Autoencoders: Select-and-Project for Fairness and Control from Encoder Features Alone
The "Sparsity vs Reconstruction Tradeoff" Illusion
Disentangling concept semantics via multilingual averaging in Sparse Autoencoders
