研究人员推出 DarwinX，这是一个采用自然选择原理来进化大型语言模型 (LLM) 智能体“束缚”的新颖系统。DarwinX 不修改模型权重，而是通过将提示、工具、技能和控制流视为经历选择的种群来优化它们。这种方法旨在提高智能体在各种基准和任务上的能力和适应性，而无需手工挑选的解决方案或黄金标准。 AI
影响 这项研究可能通过将智能体能力与其固定的模型权重脱钩，从而实现更强大、更具适应性的 AI 智能体。
排序理由 该集群包含一篇详细介绍进化 LLM 智能体束缚新方法的论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2608.07545v1 Announce Type: cross Abstract: An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops already edit harnesses, yet single-lineage search is path-dependent and local wins of…
An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops already edit harnesses, yet single-lineage search is path-dependent and local wins often regress other tasks. We introduce DarwinX, whi…
DarwinX evolves agent harnesses via population selection with frozen models, improving verified performance across benchmarks without benchmark-specific patches.