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

Multi-hop Retrieval on HotPotQA (Including Efficiency Metrics)

0.065Latency (s)

Search-R1-7B

-2.552815.1173532.787550.45765Jul 10, 2025Aug 31, 2025Oct 23, 2025Dec 14, 2025Feb 5, 2026Mar 29, 2026May 21, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2025.07
0.06548.22,1192.991.28
2025.07
0.10550.12,9371.791.77
2026.05
0.13----
2025.07
0.133569.14,4586.172.22
2025.07
0.148664.37,7701.244
2025.07
0.241570.49,1381.272.89
2025.07
0.657164.89,6571.312.75
2026.05
0.93----
2026.05
1.13----
2026.05
4.37----
2026.05
9.26----
2026.05
37.81----
2026.05
50.87----
2026.05
52.53----
2026.05
65.51----