Giters

GasolSun36 / Iter-CoT

[NAACL 2024] Enhancing Chain-of-Thoughts Prompting with Iterative Bootstrapping in Large Language Models

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Comparison with Zero-Shot-CoT

AiMijie opened this issue · comments

commented

Have you made a comparison with Zero-Shot-CoT?

commented

We didn't compare with Zero-Shot-CoT, because for mathematical reasoning, commonsense reasoning and symbolic reasoning, the effect of the Few-shot-CoT must be better than the Zero-Shot-CoT so we only did the experiments under the Few-Shot. Moreover, Iter-CoT is more focused on finding the most suitable exemplars.

commented

Got it. Thanks for your reply.