Tweeted By @_akhaliq
Large Language Models Can Self-Improve
— AK (@_akhaliq) October 24, 2022
abs: https://t.co/Hjp5AfvFD3
approach improves the general reasoning ability of a 540B-parameter LLM (74.4%→82.1% on GSM8K, 78.2%→83.0% on DROP, 90.0%→94.4% on OpenBookQA, and 63.4%→67.9% on ANLI-A3) pic.twitter.com/yYW7e2WsPB