List your company

Question Answering on HotpotQA (PR-AUC)

0.681PR-AUCCocoaPPL
0.118360.264430.41050.55657Dec 10, 2025

Evaluation results

MethodLinks
CocoaPPLStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.681
DissimilarityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.626
CocoaPPLStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.625
DissimilarityStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.611
CocoaMSPStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.523
EccentricityStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.514
EigVecDissimilarityStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.487
EccentricityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.453
EigVecDissimilarityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.448
CocoaMSPStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.43
Degree MatrixStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.311
SARStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.296
Semantic EntropyStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.278
SARStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.266
Degree MatrixStrategy=beam-search, Backbone=Gemma 3 4B base2025.120.244
Semantic EntropyStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.120.14