CRAG
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CRAG
42.3Truthfulness
36
CRAG multi-hop subset (train)
92Precision
36
CRAG
91F1 Score
36
CRAG (test)
89.4DICE Score
26
CRAG
87.4F1 Score
19
CRAG-MM (Overall)
20.5Truthfulness
18
CRAG
92.8Dice Object
17
CRAG
30.08F1 Score
15
CRAG
20.1Finance Score
12
CRAG (5-fold cross-validation)
79.6Dice Score
8
CRAG Dpath (test)
0.785Dice
8
CRAG (test)
63.3P@1
6
CRAG
88.58Dice Score
5
CRAG
16.4Finance Accuracy
5
CRAG (20% labeled data)
90.9Dice Object
3
CRAG Task 2 (test)
41Accuracy (%)
3
CRAG Task 1 (test)
34.23Accuracy
3