From Noisy Traces to Stable Gradients: Bias-Variance Optimized Preference Optimization for Aligning Large Reasoning Models

Documenti analoghi

• TGDPO: Harnessing Token-Level Reward Guidance for Enhancing Direct Preference Optimization
  di: Zhu, Mingkang, et al.
  Pubblicazione: (2025)
• Stratified GRPO: Handling Structural Heterogeneity in Reinforcement Learning of LLM Search Agents
  di: Zhu, Mingkang, et al.
  Pubblicazione: (2025)
• Enhancing LLM Knowledge Learning through Generalization
  di: Zhu, Mingkang, et al.
  Pubblicazione: (2025)
• Modular Customization of Diffusion Models via Blockwise-Parameterized Low-Rank Adaptation
  di: Zhu, Mingkang, et al.
  Pubblicazione: (2025)
• VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning
  di: Yang, Senqiao, et al.
  Pubblicazione: (2025)