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

Information Retrieval on HotpotQA (test)

69.4NDCG@10

Ours (Predictive)

57.9660.9363.966.87Nov 7, 2024Feb 7, 2025May 11, 2025Aug 12, 2025Nov 13, 2025Feb 14, 2026May 18, 2026
Updated 14d ago

Evaluation Results

MethodLinks
69.4---74.263.82.6
2024.11
68.4------
2026.05
68.2---73.162.42.8
2024.11
67.9------
2024.11
67.8------
67.5---72.461.92.9
2026.05
67.1---71.861.53.1
2024.11
66.9------
2024.11
66.7------
2026.05
66.3---70.960.83.3
2024.11
63.8------
2024.11
62.6------
2024.11
60.2------
2024.11
58.4------
2026.04
-93.6699.851.62---
2026.04
-94.4698.951.45---
2026.04
-94.1999.91.57---
2026.04
-94.7399.91.51---
2026.04
-94.7299.451.47---
2026.04
-95.4599.51.4---
2026.04
-94.899.451.46---
2026.04
-95.0199.51.45---
2026.04
-72.5287.22.68---
2026.04
-70.2584.152.63---
2026.04
-73.6188.12.64---
2026.04
-75.3689.352.56---