We explore how iterative revising a chain of thoughts with the help of information retrieval significantly improves large language models' reasoning and generation ability in long-horizon generation tasks, while hugely mitigating hallucination. In particular, the proposed method -- *retrieval-augmented thoughts* (RAT) -- revises each thought step one by one with retrieved information relevant to the task query, the current and the past thought steps, after the initial zero-shot CoT is generated. Applying RAT to GPT-3.5, GPT-4, and CodeLLaMA-7b substantially improves their performances on various long-horizon generation tasks; on average of relatively increasing rating scores by 13.63% on code generation, 16.96% on mathematical reasoning, 19.2% on creative writing, and 42.78% on embodied task planning. The demo page can be found at https://craftjarvis.github.io/RAT

翻译：我们探索如何借助信息检索迭代修正思维链，显著提升大语言模型在长程生成任务中的推理与生成能力，同时大幅缓解幻觉现象。具体而言，所提出的方法——*检索增强思维*（RAT）——在初始零样本CoT生成后，通过检索与任务查询、当前及过去思维步骤相关的信息，逐步修正每一个思维步骤。将RAT应用于GPT-3.5、GPT-4和CodeLLaMA-7b，显著提升了这些模型在多种长程生成任务上的性能；平均而言，代码生成评分相对提升13.63%，数学推理提升16.96%，创意写作提升19.2%，具身任务规划提升42.78%。演示页面见https://craftjarvis.github.io/RAT