Robust
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Robust04
46.67P@20
72
robust
46.5Recall@100
19
Robust04 Description
0.408MAP
13
Robust04 PT
50.1nDCG@10
11
Robust 2004
0.56Cohen's Kappa (κ)
11
Robust TREC 2004 (test)
51P@20
10
Robust V2 (pooled donor-holdout)
-0.293Mean Difference
9
Robust04 EN
56.38NDCG@10
8
ROBUST04 (test)
27.47AP@1000
8
Robust04 BEIR (test)
0.567nDCG@10
7
Robust04 Title queries (test)
29.04MAP
7
Robust open-world 04
34.3MAP
6
Robust04 closed-world
60.8MAP
5
Robust04 (test)
0.29MAP
5
Robust cells V2 (test)
56.7Balanced Accuracy
4
Robust cells V2 (test)
0.249Spearman Correlation (Pseudotime-depth)
2
Robust cells V2 (test)
0.867AUROC
2
Robust V2 (test)
82.8Balanced Acc
2
Robust (human evaluation)
100Control Wins
1