Recent work has shown that large language models are capable of generating
natural language reasoning steps or Chains-of-Thoughts (CoT) to answer a
multi-step question when prompted to do so. This is insufficient, however, when
the necessary knowledge is not available or up-to-date within a model's
parameters. A straightforward approach to address this is to retrieve text from
an external knowledge source using the question as a query and prepend it as
context to the model's input. This, however, is also insufficient for
multi-step QA where \textit{what to retrieve} depends on \textit{what has
already been derived}. To address this issue we propose IRCoT, a new approach
that interleaves retrieval with CoT for multi-step QA, guiding the retrieval
with CoT and in turn using retrieved results to improve CoT. Our experiments
with GPT3 show substantial improvements in retrieval (up to 22 points) and
downstream QA (up to 16 points) over the baselines on four datasets: HotpotQA,
2WikiMultihopQA, MuSiQue, and IIRC. Notably, our method also works well for
much smaller models such as T5-Flan-large (0.7B) without any additional
training.

提出了 IRCoT 方法，在多步 QA 过程中使用检索和 CoT 交替进行，同时使用检索的结果来改善 CoT。在 HotpotQA、2WikiMultihopQA、MuSiQue 和 IIRC 四个数据集上，使用 IRCoT 方法取得了显著的检索和 QA 性能改进，即便是在小型模型 T5-Flan-large（0.7B）上也能奏效。