List your company
ResearchFormerly WizWand

Memory-based Bias Reduction on Bias Reduction Benchmark Memory

91.7Bias Reduction PerformanceRDG
-947.052-677.376-407.7-138.024Nov 28, 2025
Updated 8mo ago

Evaluation results

MethodLinks
RDGModel=GPT-4o2025.1191.7
RDGModel=deepseek-r12025.1190.8
RDGModel=GPT-4o-mini2025.1183.7
RDGModel=Average2025.1181.5
ExplicitModel=GPT-4o-mini2025.1165.2
RDGModel=qwen3-235b-a22b2025.1159.7
CDDModel=deepseek-r12025.1158.2
CDDModel=GPT-4o-mini2025.1155.2
CDDModel=Average2025.1154.2
CDDModel=GPT-4o2025.1153.3
CoTModel=GPT-4o-mini2025.1150.4
CDDModel=qwen3-235b-a22b2025.1150
Self-debiasModel=GPT-4o-mini2025.1143.2
ExplicitModel=GPT-4o2025.1140
Counterfact.Model=GPT-4o-mini2025.1130.2
ExplicitModel=Average2025.1129.9
CoTModel=GPT-4o2025.1126.7
CoTModel=deepseek-r12025.1115.3
ExplicitModel=deepseek-r12025.1114.3
ExplicitModel=qwen3-235b-a22b2025.110
Self-debiasModel=deepseek-r12025.11-10.2
DDDModel=GPT-4o-mini2025.11-13.4
DDDModel=deepseek-r12025.11-18.4
DDDModel=Average2025.11-20.5
DDDModel=GPT-4o2025.11-25
DDDModel=qwen3-235b-a22b2025.11-25
Self-debiasModel=Average2025.11-68.7
Self-debiasModel=GPT-4o2025.11-75.8
Counterfact.Model=GPT-4o2025.11-96.7
CoTModel=Average2025.11-104.7
Self-debiasModel=qwen3-235b-a22b2025.11-231.9
Counterfact.Model=Average2025.11-441.3
CoTModel=qwen3-235b-a22b2025.11-511.1
Counterfact.Model=qwen3-235b-a22b2025.11-791.7
Counterfact.Model=deepseek-r12025.11-907.1