SpecEvo：多LLM推测进化实现低成本科学发现
原标题：SpecEvo: Speculative Evolution with Large Language Models for Cost-Efficient Scientific Discovery
AI 摘要
研究团队提出 SpecEvo 多 LLM 进化框架，通过轻量模型并行探索、顾问模型整合搜索证据、前沿模型在稀疏检查点进行策略干预的「推测-整合」流程，在 176 个科学发现任务上匹配或超越前沿模型基线，API 成本最高降低 80%，并在方程发现任务上提升评估效率、较最强 CO-Bench 基线提升 3.8%。
正文节选
SpecEvo: Speculative Evolution with Large Language Models for Cost-Efficient Scientific Discovery Abstract LLM-driven evolutionary search is a promising paradigm for scientific discovery, but existing methods remain costly and slow, often applying expensive reasoning uniformly while underusing evidence from failed, redundant, and saturated trials. We introduce Speculative Evolution (SpecEvo), a multi-LLM framework that coordinates lightweight and frontier models through continual consolidation o