When Reasoning Traces Become Performative: Step-Level Evidence that Chain-of-Thought Is an Imperfect Oversight Channel

Documenti analoghi

• Persona Non Grata: Single-Method Safety Evaluation Is Incomplete for Persona-Imbued LLMs
  di: Li, Wenkai, et al.
  Pubblicazione: (2026)
• Can We Trust AI Explanations? Evidence of Systematic Underreporting in Chain-of-Thought Reasoning
  di: Mehta, Deep Pankajbhai
  Pubblicazione: (2025)
• Better Accuracies, Worse Reasoning: A Step-Level Audit of Medical Chain-of-Thought Distillation
  di: Jiang, Zhaoyang, et al.
  Pubblicazione: (2026)
• User Perceptions vs. Proxy LLM Judges: Privacy and Helpfulness in LLM Responses to Privacy-Sensitive Scenarios
  di: Wu, Xiaoyuan, et al.
  Pubblicazione: (2025)
• Markov Chain of Thought for Efficient Mathematical Reasoning
  di: Yang, Wen, et al.
  Pubblicazione: (2024)