List your company
ResearchFormerly WizWand

Knowledge Retrieval on LAMA (full)

37.57AccuracyCoherence Boosting (CB)
7.316415.170723.02530.8793Oct 15, 2021
Updated 8mo ago

Evaluation results

MethodLinks
Coherence Boosting (CB)Model=GPT-3 175B, alpha_k=-0.4, k*=22021.1037.57
Coherence Boosting (CB)Model=GPT-3 13B, alpha_k=-0.5, k*=12021.1035.02
Contextual Calibration (CC)Model=GPT-3 175B2021.1034.78
Coherence Boosting (CB)Model=GPT-3 6.7B, alpha_k=-0.5, k*=12021.1032.25
Contextual Calibration (CC)Model=GPT-3 13B2021.1031.96
Contextual Calibration (CC)Model=GPT-3 6.7B2021.1030.57
Coherence Boosting (CB)Model=GPT-3 2.7B, alpha_k=-0.5, k*=12021.1028.75
f_baseModel=GPT-3 175B2021.1026.76
Contextual Calibration (CC)Model=GPT-3 2.7B2021.1026.3
Coherence Boosting (CB)Model=GPT-2 760M, alpha_k=-0.5, k*=32021.1025.74
Coherence Boosting (CB)Model=GPT-2 1.6B, alpha_k=-0.5, k*=32021.1025.43
Contextual Calibration (CC)Model=GPT-2 1.6B2021.1023.97
Contextual Calibration (CC)Model=GPT-2 760M2021.1023.9
Contextual Calibration (CC)Model=GPT-2 350M2021.1022.87
Coherence Boosting (CB)Model=GPT-2 350M, alpha_k=-0.5, k*=22021.1022.87
f_baseModel=GPT-3 13B2021.1022.06
Coherence Boosting (CB)Model=GPT-2 125M, alpha_k=-0.5, k*=12021.1019.85
f_baseModel=GPT-3 6.7B2021.1019.42
Contextual Calibration (CC)Model=GPT-2 125M2021.1017.45
f_baseModel=GPT-3 2.7B2021.1017.33
f_baseModel=GPT-2 350M2021.1014.78
f_baseModel=GPT-2 1.6B2021.1014.29
f_baseModel=GPT-2 760M2021.1013.88
f_baseModel=GPT-2 125M2021.108.48