search

allenai / reward-bench

RewardBench: the first evaluation tool for reward models.
https://huggingface.co/spaces/allenai/reward-bench
Apache License 2.0
291 stars 30 forks source link

Add New reward models #150

Closed YangRui2015 closed 1 week ago

YangRui2015 commented 1 week ago

Two new reward models are available: Ray2333/GRM-llama3-8B-distill (https://huggingface.co/Ray2333/GRM-llama3-8B-distill), Ray2333/Gemma-2B-rewardmodel-baseline (https://huggingface.co/Ray2333/Gemma-2B-rewardmodel-baseline). They are both finetuned on opensource datasets, achieving average scores of 86.1 and 73.7 locally.

Details are on the huggingface pages.

natolambert commented 1 week ago

Nice, will add these shortly with Ray2333/GRM-Gemma-2B-sftreg and Ray2333/GRM-llama3-8B-sftreg. Cool paper!

YangRui2015 commented 1 week ago

Thank you!