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

Evidence retrieval on tau-bench multi-burst-augmented cache

68Hit@2

OSU-Mem + GF

-2.7215.643452.36Jun 19, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
6838.6
2026.06
6838.6
2026.06
67.638.4
2026.06
67.336.6
2026.06
64.633.5
2026.06
57.431.8
2026.06
5228.8
2026.06
46.624.3
2026.06
46.527.1
2026.06
43.524.1
2026.06
43.524.1
2026.06
41.923.3
2026.06
38.720.9
2026.06
34.721.9
2026.06
34.720.1
2026.06
33.320.1
2026.06
32.820.3
2026.06
31.620.3
2026.06
28.515.7
2026.06
28.118
2026.06
2514.5
2026.06
2514.5
2026.06
2514.5
2026.06
19.313.3
2026.06
18.812.7
2026.06
18.112.3
2026.06
17.213.4
2026.06
17.113.6
2026.06
16.317.8
2026.06
16.117.7
2026.06
159.3
2026.06
14.58.6
2026.06
14.112.2
2026.06
1312.3
2026.06
12.58.5
2026.06
12.58.5
2026.06
12.58.5
2026.06
10.96.1
2026.06
9.98
2026.06
8.19.1
2026.06
8.18.6
2026.06
7.99
2026.06
6.34.5
2026.06
5.17.6
2026.06
3.77.9
2026.06
3.27.6
2026.06
2.65.3
2026.06
0.54.9
2026.06
0.33.5
2026.06
0.33.4
2026.06
02.6
2026.06
04.4