RewardBench: Evaluating Reward Models for Language Modeling

Explore score history ↗

RewardBench: Evaluating Reward Models for Language Modeling is an AI benchmark in the Benchmark Radar catalog.

Released: 2024-03-21
Openness: unknown
Importer: Benchmark Saturation

Source provenance

Original evidence

RewardBench: Evaluating Reward Models for Language Modeling results and reported scores

Reported benchmark scores · Benchmark Saturation

Reported scores: 5

Model	Organization	Reported value	Reported	Evidence
infly/INF-ORM-Llama3.1-70B	—	95.1	—	evidence
nicolinho/QRM-Gemma-2-27B	—	94.4	—	evidence
nvidia/Llama-3.1-Nemotron-70B-Reward	—	94.1	—	evidence
ShikaiChen/LDL-Reward-Gemma-2-27B-v0.1	—	95.0	—	evidence
Skywork/Skywork-Reward-Gemma-2-27B-v0.2	—	94.3	—	evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite RewardBench: Evaluating Reward Models for Language Modeling?