List your company
ResearchFormerly WizWand

Reward Modeling on RewardBench latest (test)

74.9AccuracyTwo-anchor-Skywork
43.49251.64659.867.954May 12, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Two-anchor-SkyworkBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0574.90.19970.5898
BT(Hard)Backbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0573.50.19830.5822
Two-anchor-DeepSeek-ProBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0573.50.2060.6017
Two-anchor-DeepSeek-FlashBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0572.70.19830.6085
BTBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0572.20.20380.5958
Two-anchor-SkyworkBackbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0571.70.20990.636
Two-anchor-DeepSeek-ProBackbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0571.70.21550.6301
Two-anchor-SkyworkBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.0571.60.20460.7231
Two-anchor-DeepSeek-FlashBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.0571.50.2020.6308
BTBackbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0571.30.23930.7217
GaussianBackbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0570.60.23640.7153
Two-anchor-DeepSeek-FlashBackbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0570.30.22250.6508
Two-anchor-DeepSeek-ProBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.05700.21970.7331
BTBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.0569.10.22170.8144
GaussianBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer2-9K2026.0568.70.21920.7195
BT(Hard)Backbone=Llama-3.2-1B-Instruct, Training Dataset=PersonalLLM-40K2026.0568.60.27031.2759
Two-anchor-SkyworkBackbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.05670.21640.6263
GaussianBackbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.0565.70.23660.9594
BT(Hard)Backbone=Llama-3.2-1B-Instruct, Training Dataset=HelpSteer3-20K2026.0564.80.25180.8945
BT(Hard)Backbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.0558.60.30961.1299
Two-anchor-DeepSeek-FlashBackbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.0558.60.24090.6808
Two-anchor-DeepSeek-ProBackbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.0555.60.23970.6728
GaussianBackbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.0548.10.26440.7294
BTBackbone=Llama-3.2-1B-Instruct, Training Dataset=MultiPref-8K2026.0544.70.26350.7263