Reward Modeling on PKU-SafeRLHF (test)
0.074MAESelectiveRM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SelectiveRMMethod Category=Ours2026.05 | 0.074 | 0.057 | 0.772 | — | |
| ϵ-SoftmaxMethod Category=Heuristic Methods2026.05 | 0.078 | 0.067 | 0.731 | — | |
| ILDELearning Category=PU learning methods2026.03 | 0.0871 | — | 0.7104 | 0.2681 | |
| CCRMethod Category=Statistically Consistent Methods2026.05 | 0.1 | 0.073 | 0.707 | — | |
| ImplicitRM2026.03 | 0.1053 | — | 0.7872 | 0.2294 | |
| F-correctionMethod Category=Statistically Consistent Methods2026.05 | 0.113 | 0.083 | 0.668 | — | |
| CausalRM-IPSScenario=1, rho01=0.2, rho10=0.12026.03 | 0.119 | 0.057 | 0.77 | — | |
| UBPRLearning Category=PU learning methods2026.03 | 0.1251 | — | 0.628 | 0.3039 | |
| NNPULearning Category=PU learning methods2026.03 | 0.1279 | — | 0.6471 | 0.296 | |
| UPLLearning Category=PU learning methods2026.03 | 0.1282 | — | 0.6742 | 0.2844 | |
| CausalRM-DRScenario=1, rho01=0.2, rho10=0.12026.03 | 0.129 | 0.055 | 0.779 | — | |
| UPULearning Category=PU learning methods2026.03 | 0.129 | — | 0.6043 | 0.3134 | |
| SelectMixLearning Category=PU learning methods2026.03 | 0.1422 | — | 0.7346 | 0.2567 | |
| CUBPRLearning Category=PU learning methods2026.03 | 0.1423 | — | 0.6389 | 0.2994 | |
| LAGAMLearning Category=PU learning methods2026.03 | 0.1465 | — | 0.754 | 0.2472 | |
| CSGNMethod Category=Statistically Consistent Methods2026.05 | 0.155 | 0.065 | 0.738 | — | |
| CausalRM-DRScenario=2, rho01=0.1, rho10=0.22026.03 | 0.171 | 0.07 | 0.72 | — | |
| SDRLearning Category=Debiased learning methods2026.03 | 0.1771 | — | 0.6948 | 0.2753 | |
| DRLearning Category=Debiased learning methods2026.03 | 0.1871 | — | 0.5011 | 0.352 | |
| MTIPSLearning Category=Debiased learning methods2026.03 | 0.1899 | — | 0.7093 | 0.2687 | |
| ILDEMethod Category=Heuristic Methods2026.05 | 0.19 | 0.062 | 0.752 | — | |
| MTDRLearning Category=Debiased learning methods2026.03 | 0.1959 | — | 0.6876 | 0.2785 | |
| SelectMixMethod Category=Heuristic Methods2026.05 | 0.199 | 0.064 | 0.741 | — | |
| Robust DivideMixMethod Category=Heuristic Methods2026.05 | 0.205 | 0.067 | 0.731 | — | |
| LabelWaveMethod Category=Heuristic Methods2026.05 | 0.216 | 0.067 | 0.729 | — | |
| kMEIDTMMethod Category=Statistically Consistent Methods2026.05 | 0.223 | 0.072 | 0.712 | — | |
| NLSMethod Category=Heuristic Methods2026.05 | 0.236 | 0.074 | 0.703 | — | |
| ROBOTMethod Category=Statistically Consistent Methods2026.05 | 0.24 | 0.074 | 0.703 | — | |
| IPSLearning Category=Debiased learning methods2026.03 | 0.2423 | — | 0.5914 | 0.2718 | |
| CNLCUMethod Category=Heuristic Methods2026.05 | 0.243 | 0.081 | 0.673 | — | |
| Co-TeachingMethod Category=Heuristic Methods2026.05 | 0.245 | 0.082 | 0.671 | — | |
| HOCMethod Category=Statistically Consistent Methods2026.05 | 0.253 | 0.079 | 0.683 | — | |
| CDRMethod Category=Heuristic Methods2026.05 | 0.255 | 0.079 | 0.682 | — | |
| NaiveMethod Category=Statistically Consistent Methods2026.05 | 0.273 | 0.089 | 0.642 | — | |
| NaiveLearning Category=Debiased learning methods2026.03 | 0.2887 | — | 0.5535 | 0.333 | |
| BPRLearning Category=PU learning methods2026.03 | 0.3115 | — | 0.5228 | 0.3442 |