paper-buf

• LLM safety
  • Towards scalable oversight: Meta-evaluation of llms as evaluators via agent debate, the arxiv version, and code
  • Transformer feed-forward layers build predictions by promoting concepts in the vocabulary space
  • NOTES
• LLM
  • Evaluating the ripple effects of knowledge editing in language models
  • NOTES
  • From understanding to utilization: A survey on explainability for large language models
  • A practical review of mechanistic interpretability for transformer-based language models