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

Document Question Answering on DUDE (Accuracy Metrics)

80.7Accuracy (all)

GPT-4o-200b-128

12.37230.11147.8565.589Oct 8, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.10
80.78363.3
2025.10
72.574.654.5
2025.10
727554.1
2025.10
70.773.253.4
2025.10
70.673.253.5
2025.10
64.169.757.4
2025.10
6366.953.3
2025.10
60.263.646.7
2025.10
59.665.646.7
2025.10
58.960.941.6
2025.10
58.265.646.6
2025.10
57.863.844.6
2025.10
57.659.135
2025.10
57.664.345
2025.10
56.261.542.2
2025.10
5656.436.7
2025.10
54.557.843.3
2025.10
53.557.442.6
2025.10
52.95641.7
2025.10
51.451.241.7
2025.10
192115
2025.10
1720.513
2025.10
1518.811.6