Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-Hop Question Answering on 2Wiki (EM, FBERT)
Loading...
47.91
EM
CoT
20.7764
27.8207
34.865
41.9093
Apr 20, 2026
EM
FBERT
Updated 1mo ago
Evaluation Results
Method
Method
Links
EM
FBERT
CoT
SLM=Qwen3-235B
2026.04
47.91
76.81
Dynamic Collaboration Framework
SLM=Qwen3-8B
2026.04
47.09
77.35
Direct Infer
SLM=Qwen3-235B
2026.04
41.57
72.4
Dynamic Collaboration Framework
SLM=Qwen3-4B
2026.04
40.52
74.33
Dynamic Collaboration Framework
SLM=Llama-3.1-8B
2026.04
39.27
73.23
Dynamic Collaboration Framework
SLM=Llama-3.2-3B
2026.04
38.33
72.3
CoT
SLM=Qwen3-8B
2026.04
33.51
68.65
CoT
SLM=Qwen3-4B
2026.04
32.2
68.26
CoT
SLM=Llama-3.1-8B
2026.04
22.44
58.69
CoT
SLM=Llama-3.2-3B
2026.04
21.82
58.63
Feedback
Search any
task
Search any
task