Finance
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Finance
29.19ROUGE-Lsum
8
Finance (private)
91.98TEDS (Text)
7
Finance
62.3Label Quality
5
Finance
0.62Hit@1
4
Finance synthetic rewrites
0Epsilon (ε)
4
Finance
17,695Output Size
4
Finance Out-of-domain
42.4ROUGE-1
4
Finance Out-of-domain
55.8Accuracy
4
Finance
97Win Rate
3
Finance (test)
99.3Accuracy
3
Finance v3 v4 (Unsafe vs. Safe Holdout)
99.2AUROC
2
Finance (Semi-synthetic)
67AUCadj
2
Finance Extrapolation
25.9TtT0.2
2
Finance Interpolation
45TtT0.2
2
Finance
66Accuracy
2
Finance instantiation
173Matched Pairs
1