Measuring Five-Nines Reliability: Sample-Efficient LLM Evaluation in Saturated Benchmarks

Documenti analoghi

• Test-Time Adaptation Induces Stronger Accuracy and Agreement-on-the-Line
  di: Kim, Eungyeup, et al.
  Pubblicazione: (2023)
• Not All Rollouts are Useful: Down-Sampling Rollouts in LLM Reinforcement Learning
  di: Xu, Yixuan Even, et al.
  Pubblicazione: (2025)
• AcceleratedLiNGAM: Learning Causal DAGs at the speed of GPUs
  di: Akinwande, Victor, et al.
  Pubblicazione: (2024)
• FUSE-ing Language Models: Zero-Shot Adapter Discovery for Prompt Optimization Across Tokenizers
  di: Williams, Joshua Nathaniel, et al.
  Pubblicazione: (2024)
• Equilibrium Reasoners: Learning Attractors Enables Scalable Reasoning
  di: Huang, Benhao, et al.
  Pubblicazione: (2026)