Repository
chain-of-thought-hub
Benchmarking large language models' complex reasoning ability with chain-of-thought prompting
Saved on this device. Sign in to sync across devices.
Repository
Benchmarking large language models' complex reasoning ability with chain-of-thought prompting