Google New Research: LEAP Framework Enables Efficient Solving of Mathematical Proofs with General LLMs
elvis(@omarsar0)144 字 (约 1 分钟)
88
Google's LEAP framework wraps a general-purpose LLM in an agentic scaffold that grounds every step in the Lean compiler and iterates against verifier feedback. It solves all 12 Putnam 2025 problems with one model, lifting the one-shot solve rate of the Lean-IMO-Bench from under 10% to 70%, outperforming a specialized gold-medal system that scores 48. Paper: arXiv:2606.03303. Learn to build effective AI agents at academy.dair.ai.
入选理由：LEAP 通用 LLM 一模型解决全部 12 道 Putnam 2025 题。
FeaturedTweet#LEAP#Lean compiler#Putnam 2025#agentic framework#general-purpose LLM英文