Who & When
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Who&When
8.1Pair µF1
30
Who&When
60.79Agent Accuracy
22
WHO&WHEN Algorithm-Generated
38.9Accuracy@0
19
WHO&WHEN Human-Crafted
17.2Acc@0
19
Who&When Hand (58 traces)
31Top Accuracy
16
Who&When Algo (126 traces)
49.2Accuracy
16
Who&When (HC)
22.99Accuracy (%)
15
Who&When (Alg)
42.86Accuracy (Who&When Alg)
15
Who&When n=58 (Hand-Crafted)
73Agent-level Accuracy
15
Who&When Algorithm-Generated n=126
68Agent-level Accuracy
15
Who&When Total
36.22Step-level Accuracy
13
Who&When Hand-Crafted
41.38Step-level Accuracy
13
Who&When Algorithm-Generated
42.86Step-level Accuracy
13
Who&When
57.69Step Accuracy
8
Who&When
100Eta (%)
6
Who & When Boundary
38.71Agent Attribution Accuracy
6
Who & When Remove ID
26.47Agent Attribution Accuracy
6
Who & When Baseline
54.33Agent Attribution Accuracy
6
Who&When HC
21.26Top-1 Accuracy
5
Who&When (Alg)
28.84Top-1 Accuracy
5
Who&When attribution benchmark
60Rec-Exact
4
Who&When HC
79,817.71Input Tokens
4
Who&When (Alg)
15,673.79Input Tokens
4
Who&When
21.1Top-1 Accuracy
3