Your Language Model May Think Too Rigidly: Achieving Reasoning Consistency with Symmetry-Enhanced Training

Documenti analoghi

• Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety
  di: Zhang, Yuyou, et al.
  Pubblicazione: (2025)
• Feasibility Consistent Representation Learning for Safe Reinforcement Learning
  di: Cen, Zhepeng, et al.
  Pubblicazione: (2024)
• Behavior Injection: Preparing Language Models for Reinforcement Learning
  di: Cen, Zhepeng, et al.
  Pubblicazione: (2025)
• Learning from Sparse Offline Datasets via Conservative Density Estimation
  di: Cen, Zhepeng, et al.
  Pubblicazione: (2024)
• Constraint-Conditioned Policy Optimization for Versatile Safe Reinforcement Learning
  di: Yao, Yihang, et al.
  Pubblicazione: (2023)