2024

• AI Safety: A Climb To Armageddon? 11/13
• Interpreting Attention Layer Outputs with Sparse Autoencoders 11/13
• Mechanistically Interpreting a Transformer-based 2-SAT Solver: An Axiomatic Approach 11/13
• Safety Cases: How to Justify the Safety of Advanced AI Systems 11/13
• Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress? 11/13
• The Elephant in the Room - Why AI Safety Demands Diverse Teams 11/13
• The Role of AI Safety Institutes in Contributing to International Standards for Frontier AI Safety 11/13
• Bilinear MLPs enable weight-based mechanistic interpretability 11/11
• Dictionary Learning Improves Patch-Free Circuit Discovery in Mechanistic Interpretability: A Case Study on Othello-GPT 11/11
• Extracting Finite State Machines from Transformers 11/11
• How to use and interpret activation patching 11/11
• Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning 11/11
• Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoders 11/11
• Mechanistic Interpretability for AI Safety - A Review 11/11
• Mechanistic Interpretability of Reinforcement Learning Agents 11/11
• Mechanistic Understanding and Mitigation of Language Model Non-Factual Hallucinations 11/11
• Mechanistic Unlearning: Robust Knowledge Unlearning and Editing via Mechanistic Localization 11/11
• Scalable Mechanistic Neural Networks 11/11
• Transcoders Find Interpretable LLM Feature Circuits 11/11
• Unlocking the Future: Exploring Look-Ahead Planning Mechanistic Interpretability in Large Language Models 11/11
• Using Degeneracy in the Loss Landscape for Mechanistic Interpretability 11/11