List your company

Question Answering on PIQA (test)

85AccuracyExpanded Drop
23.920839.777955.63571.4921Mar 7, 2025May 14, 2025Jul 21, 2025Sep 27, 2025Dec 4, 2025Feb 10, 2026Apr 19, 2026

Evaluation results

MethodLinks
Expanded DropModel=Mixtral-8x7B-Instruct, Capacity Factor=1.52025.0385
BaselineModel=Mixtral-8x7B-Instruct, Capacity Factor=1.52025.0384.8
Token DropModel=Mixtral-8x7B-Instruct, Capacity Factor=1.52025.0383.3
Expert DropModel=Mixtral-8x7B-Instruct, Capacity Factor=1.52025.0383.2
BaselineModel=DeepSeek-V2-Lite-Chat, Capacity Factor=2.02025.0381.4
BaselineModel=OLMoE-Instruct, Capacity Factor=2.02025.0380.2
BaselineModel=Qwen1.5-MoE-Chat, Capacity Factor=1.52025.0379.9
Expanded DropModel=OLMoE-Instruct, Capacity Factor=2.02025.0379.4
Expanded DropModel=DeepSeek-V2-Lite-Chat, Capacity Factor=2.02025.0379.4
Expanded DropModel=Qwen1.5-MoE-Chat, Capacity Factor=1.52025.0379.1
Token DropModel=Qwen1.5-MoE-Chat, Capacity Factor=1.52025.0378.8
Expert DropModel=Qwen1.5-MoE-Chat, Capacity Factor=1.52025.0378.7
Token DropModel=DeepSeek-V2-Lite-Chat, Capacity Factor=2.02025.0378.3
Token DropModel=OLMoE-Instruct, Capacity Factor=2.02025.0377.9
OracleModel Size=Large2026.0477.8
OracleModel Size=Large, Noise Type=Uniform2026.0477.8
REALMModel Size=Large, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0477.67
Expert DropModel=DeepSeek-V2-Lite-Chat, Capacity Factor=2.02025.0377.6
Expert DropModel=OLMoE-Instruct, Capacity Factor=2.02025.0376.9
REALMModel Size=Large, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0476.5
REALMModel Size=Large, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0476.5
REALMModel Size=Large, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0476.45
NOISYModel Size=Large, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0475.1
OracleModel Size=Base, Noise Type=Uniform2026.0467.68
REALMModel Size=Base, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0466.55
REALMModel Size=Base, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0465.79
REALMModel Size=Base, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0465.7
NOISYModel Size=Base, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0463.33
OracleModel Size=Small, Noise Type=Uniform2026.0452.88
REALMModel Size=Small, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0452.88
REALMModel Size=Small, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0452.14
noisyModel Size=Large, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0451.99
NOISYModel Size=Large, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0451.99
REALMModel Size=Small, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0451.6
NOISYModel Size=Base, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0451.47
NOISYModel Size=Small, Expertise Distribution=Dist. 2, Noise Type=Uniform2026.0451.12
NOISYModel Size=Small, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0451.08
NOISYModel Size=Small, Expertise Distribution=Dist. 3, Noise Type=Uniform2026.0450.36
NOISYModel Size=Base, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0446.31
NOISYModel Size=Large, Expertise Distribution=Dist. 1, Noise Type=Uniform2026.0426.27