Do We Truly Need So Many Samples? Multi-LLM Repeated Sampling Efficiently Scales Test-Time Compute

Documenti analoghi

• Stop Overvaluing Multi-Agent Debate -- We Must Rethink Evaluation and Embrace Model Heterogeneity
  di: Zhang, Hangfan, et al.
  Pubblicazione: (2025)
• Beyond GPT-5: Making LLMs Cheaper and Better via Performance-Efficiency Optimized Routing
  di: Zhang, Yiqun, et al.
  Pubblicazione: (2025)
• Provably Efficient Information-Directed Sampling Algorithms for Multi-Agent Reinforcement Learning
  di: Zhang, Qiaosheng, et al.
  Pubblicazione: (2024)
• Graph Attention is Not Always Beneficial: A Theoretical Analysis of Graph Attention Mechanisms via Contextual Stochastic Block Models
  di: Ma, Zhongtian, et al.
  Pubblicazione: (2024)
• Learning Compact Representations of LLM Abilities via Item Response Theory
  di: Chen, Jianhao, et al.
  Pubblicazione: (2025)