让大模型通过协作验证实现知识积累，突破静态训练瓶颈。
The Ratchet Effect in Silico: How Interaction Drives Cumulative Intelligence in Large Language Models
- 构建异质智能体群体，通过互相验证与共享记忆实现知识迭代。
- 1-40亿参数模型在数学推理上提升8.8至18.9分，接近700亿单体模型表现。
- 强调同行验证是知识累积的核心机制，适合研究智能演化与协作系统者。
人类智能通过累积文化进化（CCE）实现增长，这是一种将创新保留对抗熵增的机制。相比之下，大语言模型训练仍主要依赖静态语料和参数膨胀，缺乏通过交互实现内在积累的空间。本文提出POLIS（群体协同学习与推理社会）框架：异质智能体生成解决方案，相互验证输出，将有效成果存入共享文化记忆，并通过参数更新内化知识。在数学推理基准测试中，1–40亿参数模型群体相比基础模型平均提升8.8–18.9分，缩小了与700亿以上单体模型的差距。机制消融实验表明，同行验证是核心的‘锁扣’机制，内化过程可维持多轮知识积累，为认知警惕性推动持久知识增长提供了计算证据。结果表明，结构化社交互动是独立于参数规模的可扩展杠杆。
原文摘要 · Abstract (English)
Human intelligence scales through cumulative cultural evolution (CCE), a ratchet process in which innovations are retained against entropic drift. Large language model training, by contrast, still depends primarily on static corpora and parameter growth, leaving little room for endogenous accumulation through interaction. We present POLIS (Population Orchestrated Learning and Inference Society), a framework in which heterogeneous agents generate solutions, verify one another's outputs, retain validated artifacts in shared cultural memory, and internalize them through parameter updates. On mathematical reasoning benchmarks, populations of 1--4B-parameter models achieved average gains of 8.8--18.9 points over base models and narrowed the gap to 70B+ monoliths. Mechanistic ablations identify peer verification as the main ratchet operator and show that internalization sustains accumulation across rounds, providing computational evidence that epistemic vigilance organizes durable knowledge growth. These results position structured social interaction as a scaling lever orthogonal to parameter count.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。