Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Question Answering on OpenBookQA (Accuracy, TFLOPs)

93.6Accuracy

dense alignment

-1.45623.22247.972.578Jun 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
93.621.54
2026.06
93.421.79
2026.06
92.867.08
2026.06
92.615.64
2026.06
9256.24
2026.06
91.831.16
2026.06
91.831.16
2026.06
91.219.85
2026.06
91.212.56
2026.06
91.219.85
2026.06
90.437.73
2026.06
90.255.5
2026.06
906.56
2026.06
89.67.95
2026.06
88.89.42
2026.06
88.810.18
2026.06
8813.28
2026.06
889.18
2026.06
889.18
2026.06
876.64
2026.06
8733.91
2026.06
877.42
2026.06
85.64.5
2026.06
85.67.83
2026.06
85.48.9
2026.06
8543.48
2026.06
83.815.85
2026.06
82.417.15
2026.06
77.23.43
2026.06
76.45.01
2026.06
27.472.06
2026.06
27.260.83
2026.06
2714.94
2026.06
26.253.8
2026.06
2437.86
2026.06
23.232.75
2026.06
20.241.95
2026.06
16.420.05
2026.06
1014.76
2026.06
7.69.84
2026.06
6.617.93
2026.06
2.224.77