stair-lab/reeval

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 models: Llama-3.1-8B-Instruct and Mistral-7B-Instruct-v0.3, as in the embed_meta-llama_Llama-3.1-8B-Instruct.pkl and embed_mistralai_Mistral-7B-Instruct-v0.3.pkl. The files in folder… See the full description on the dataset page: https://huggingface.co/datasets/stair-lab/reeval.

<iframe src="https://silytics.si/embed/item/71e6c8b3-c66e-4615-9c47-428f976d89ee" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.