A Principled Loss Function for Direct Language Model Alignment

Documenti analoghi

• Provable Scaling Laws of Feature Emergence from Learning Dynamics of Grokking
  di: Tian, Yuandong
  Pubblicazione: (2025)
• Param$Δ$ for Direct Weight Mixing: Post-Train Large Language Model at Zero Cost
  di: Cao, Sheng, et al.
  Pubblicazione: (2025)
• Sail into the Headwind: Alignment via Robust Rewards and Dynamic Labels against Reward Hacking
  di: Rashidinejad, Paria, et al.
  Pubblicazione: (2024)
• From Coefficients to Directions: Rethinking Model Merging with Directional Alignment
  di: Chen, Zhikang, et al.
  Pubblicazione: (2025)
• Diffusion World Model: Future Modeling Beyond Step-by-Step Rollout for Offline Reinforcement Learning
  di: Ding, Zihan, et al.
  Pubblicazione: (2024)