PACR: Progressively Ascending Confidence Reward for LLM Reasoning

Documenti analoghi

• PDCR: Perception-Decomposed Confidence Reward for Vision-Language Reasoning
  di: Yoon, Hee Suk, et al.
  Pubblicazione: (2026)
• Decomposed On-Policy Distillation for Vision-Language Reasoning: Steering Gradients for Visual Grounding
  di: Yoon, Hee Suk, et al.
  Pubblicazione: (2026)
• AdaMER-CTC: Connectionist Temporal Classification with Adaptive Maximum Entropy Regularization for Automatic Speech Recognition
  di: Eom, SooHwan, et al.
  Pubblicazione: (2024)
• High-Fidelity Text-to-Image Generation from Pre-Trained Vision-Language Models via Distribution-Conditioned Diffusion Decoding
  di: Hong, Ji Woo, et al.
  Pubblicazione: (2026)
• TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback
  di: Yoon, Eunseop, et al.
  Pubblicazione: (2024)