Program Distillation: Transferring LLM Capability from System 2 to System 0
arXiv:2611.19461v1 Announce Type: new Abstract: Recent studies leverage large language models (LLMs) to construct executable programs through environment interaction, feedback analysis, and code revision. However, prior work has mainly treated LLMs as optimizers, overlooking their ability to transfer knowledge across system levels. We propose Program Distillation (PD), a framework that formalizes the transfer of abilities from System 2 to System 0. During learning, the LLM provides task knowledg
阅读原始来源来源记录
项目分析
暂无五维分析,可先查看上方来源资料。