Small Vectors, Big Effects: A Mechanistic Study of RL-Induced Reasoning via Steering Vectors

Documenti analoghi

• Steering LLM Reasoning Through Bias-Only Adaptation
  di: Sinii, Viacheslav, et al.
  Pubblicazione: (2025)
• You Do Not Fully Utilize Transformer's Representation Capacity
  di: Gerasimov, Gleb, et al.
  Pubblicazione: (2025)
• Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy
  di: Balagansky, Nikita, et al.
  Pubblicazione: (2025)
• Kronecker Factorization Improves Efficiency and Interpretability of Sparse Autoencoders
  di: Kurochkin, Vadim, et al.
  Pubblicazione: (2025)
• Analyze Feature Flow to Enhance Interpretation and Steering in Language Models
  di: Laptev, Daniil, et al.
  Pubblicazione: (2025)