⭐️ We are honored that Re2 is added into optillm, a repo to optimize LLMs' inference.
Intro: The official code for the paper EMNLP 2024 "Re-reading improves reasoning in large language models".
Idea: Simply repeating the question to get bidirectional understanding for improving reasoning.
Code will be availabel in the following two weeks . Star it and Stay tuned :) Or you can simply implement it because it is very simple.