Safe Transformer: An Explicit Safety Bit For Interpretable And Controllable Alignment

Documenti analoghi

• Towards Empirical Interpretation of Internal Circuits and Properties in Grokked Transformers on Modular Polynomials
  di: Furuta, Hiroki, et al.
  Pubblicazione: (2024)
• Understanding Emergent Misalignment via Feature Superposition Geometry
  di: Minegishi, Gouki, et al.
  Pubblicazione: (2026)
• Zipping the Thought: When and How Compressed Reasoning Data Works in LLM Post-Training
  di: Matsutani, Kohsei, et al.
  Pubblicazione: (2026)
• Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words
  di: Minegishi, Gouki, et al.
  Pubblicazione: (2025)
• Inconsistent Tokenizations Cause Language Models to be Perplexed by Japanese Grammar
  di: Gambardella, Andrew, et al.
  Pubblicazione: (2025)