stair-lab/reeval download history

stair-lab/reeval is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 281 times (31 in the last 7 days), and 3,217 times in total. It ranks #51,683 among datasets by monthly downloads.

This repository implements the paper Reliable and Efficient Amortized Model-based Evaluation. We transfer the original HELM data into long format, as in the long.pkl, which is then transferred into the response matrix, as in resmat.pkl. We then obtain the embedding of each question from two language

Open stair-lab/reeval on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.