ConfPO: Exploiting Policy Model Confidence for Critical Token Selection in Preference Optimization

Documenti analoghi

• Can Video LLMs Refuse to Answer? Alignment for Answerability in Video Large Language Models
  di: Yoon, Eunseop, et al.
  Pubblicazione: (2025)
• LI-TTA: Language Informed Test-Time Adaptation for Automatic Speech Recognition
  di: Yoon, Eunseop, et al.
  Pubblicazione: (2024)
• TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback
  di: Yoon, Eunseop, et al.
  Pubblicazione: (2024)
• PACR: Progressively Ascending Confidence Reward for LLM Reasoning
  di: Yoon, Eunseop, et al.
  Pubblicazione: (2025)
• PDCR: Perception-Decomposed Confidence Reward for Vision-Language Reasoning
  di: Yoon, Hee Suk, et al.
  Pubblicazione: (2026)