LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scaling

Documenti analoghi

• ATLAS: Agentic Test-time Learning-to-Allocate Scaling
  di: Qin, Peijia, et al.
  Pubblicazione: (2026)
• DAJ: Data-Reweighted LLM Judge for Test-Time Scaling in Code Generation
  di: Qin, Peijia, et al.
  Pubblicazione: (2026)
• Models Under SCOPE: Scalable and Controllable Routing via Pre-hoc Reasoning
  di: Cao, Qi, et al.
  Pubblicazione: (2026)
• BiDoRA: Bi-level Optimization-Based Weight-Decomposed Low-Rank Adaptation
  di: Qin, Peijia, et al.
  Pubblicazione: (2024)
• DreamPRM-Code: Function-as-Step Process Reward Model with Label Correction for LLM Coding
  di: Zhang, Ruiyi, et al.
  Pubblicazione: (2025)