Sloth: scaling laws for LLM skills to predict multi-benchmark performance across families

Documenti analoghi

• tinyBenchmarks: evaluating LLMs with fewer examples
  di: Polo, Felipe Maia, et al.
  Pubblicazione: (2024)
• Efficient multi-prompt evaluation of LLMs
  di: Polo, Felipe Maia, et al.
  Pubblicazione: (2024)
• Weak Supervision Performance Evaluation via Partial Identification
  di: Polo, Felipe Maia, et al.
  Pubblicazione: (2023)
• A transfer learning framework for weak-to-strong generalization
  di: Somerstep, Seamus, et al.
  Pubblicazione: (2024)
• Bridging Human and LLM Judgments: Understanding and Narrowing the Gap
  di: Polo, Felipe Maia, et al.
  Pubblicazione: (2025)