GEdit-Bench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GEdit-Bench-CN (test)
7.898Q_SC Score
8
GEdit-Bench Chinese (CN) setting
7.92G_SC Score
7
GEdit-Bench
6.992Style Score
7
GEdit-Bench Intersection
7.819Q_SC
7
GEdit-Bench-EN (test)
8.01SC Score
7
GEdit-Bench
0.274Editing Reward
6
GEdit-Bench
63.19SC Preference
4
GEdit-Bench-EN (Full set)
7.03BC Score
4
GEdit-Bench Full set
7.134UP
4
GEdit-Bench Intersection subset (IS)
7.109UP
4
Gedit-Bench
7.06Background Change Score
3
GEdit-Bench
76Accuracy
2
GEdit-Bench (test)
7.8SC
2