Can the Inference Logic of Large Language Models be Disentangled into Symbolic Concepts?

Add code
Apr 03, 2023

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: