Re-Reading Improves Reasoning in Language Models
Primary Area: generative models
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Large Language Model, Reasoning
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
TL;DR: A straightforward, plug-and-play, effective, and universally applicable reasoning approach for LLMs—namely, re-reading the question—enhancing bidirectional comprehension of questions within the context of decoder-only causal language models.
Abstract: Reasoning presents a significant and challenging issue for Large Language Models (LLMs). The predominant focus of research has revolved around developing diverse prompting strategies to guide and structure the reasoning processes of LLMs. However, these approaches based on decoder-only causal language models often operate the input question in a single forward pass, potentially missing the rich, back-and-forth interactions inherent in human reasoning. Scant attention has been paid to a critical dimension, i.e., the input question itself embedded within the prompts. In response, we introduce a seemingly straightforward yet remarkably effective prompting strategy—Re2, which involves re-reading the question. Drawing inspiration from human learning and problem-solving, re-reading entails revisiting the question information embedded within input prompts. This approach aligns seamlessly with the cognitive principle of reinforcement, enabling LLMs to understand the input in a ''bidirectional'' manner, extract deeper insights, and ultimately enhance their reasoning capabilities across various tasks. Experiments conducted on a series of reasoning benchmarks serve to underscore the effectiveness {and generality} of our method. Moreover, our findings demonstrate that our approach seamlessly integrates with various language models, {thought-eliciting} prompting methods, and ensemble techniques, further underscoring its versatility and compatibility in the realm of LLMs.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 1103
Loading