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

Long-context Recall on SQuAD, FDA, and SWDE Subset

72.3SQuAD Recall Accuracy

Qwen3

18.63632.56846.560.432Jun 29, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
72.378.290.880.4
2026.06
58.78190.876.8
2026.06
54.568.280.967.9
2026.06
54.374.182.270.2
2026.06
54.269.481.968.5
2026.06
54.173.581.269.6
2026.06
52.867.982.467.7
2026.06
52.773.587.371.2
2026.06
51.364.780.965.7
2026.06
51.370.582.568.1
2026.06
51.171.681.668.1
2026.06
49.763.384.165.7
2026.06
48.259.278.962.1
2026.06
47.373.782.367.7
2026.06
44.182.180.568.9
2026.06
44.153.578.658.7
2026.06
44.182.180.568.9
2026.06
43.741.474.253.1
2026.06
43.657.381.260.7
2026.06
43.563.581.562.8
2026.06
41.959.780.960.8
2026.06
41.762.476.260.1
2026.06
40.264.580.461.7
2026.06
39.87985.167.9
2026.06
39.87985.167.9
2026.06
39.870.281.463.8
2026.06
39.549.474.854.5
2026.06
3956.678.658.1
2026.06
38.567.880.762.3
2026.06
38.471.376.762.1
2026.06
36.468.575.460.1
2026.06
35.470.77660.7
2026.06
35.358.276.856.7
2026.06
34.760.472.655.9
2026.06
33.572.875.660.6
2026.06
32.761.475.256.4
2026.06
31.155.67553.9
2026.06
30.351.371.851.1
2026.06
30.355.572.252.7
2026.06
29.358.570.552.8
2026.06
26.748.873.949.8
2026.06
26.56271.253.2
2026.06
25.758.470.851.6
2026.06
2335.976.845.2
2026.06
22.862.873.553
2026.06
21.453.481.552.1
2026.06
20.8417746.3
2026.06
20.752.670.648