When Should a Language Model Trust Itself? Same-Model Self-Verification as a Conditional Confidence Signal

Documenti analoghi

• Trust in One Round: Confidence Estimation for Large Language Models via Structural Signals
  di: Yang, Pengyue, et al.
  Pubblicazione: (2026)
• Self-Training Large Language Models with Confident Reasoning
  di: Jang, Hyosoon, et al.
  Pubblicazione: (2025)
• When Should Models Change Their Minds? Contextual Belief Management in Large Language Models
  di: Xu, Haoming, et al.
  Pubblicazione: (2026)
• Self-ensemble: Mitigating Confidence Mis-calibration for Large Language Models
  di: Xu, Zicheng, et al.
  Pubblicazione: (2025)
• Confidence in the Reasoning of Large Language Models
  di: Pawitan, Yudi, et al.
  Pubblicazione: (2024)