Which OpenAssistant model was used for "Human & GPT-4 Evaluations"?

#69
by tju01 - opened

In your announcement blog post, the OpenAssistant link goes to https://huggingface.co/OpenAssistant/oasst-sft-1-pythia-12b. However, on the leaderboard, oasst-12b links to https://huggingface.co/OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5. Which one is the correct one and could you please fix either the blog post or the leaderboard?

natolambert changed discussion status to closed

Sign up or log in to comment