Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-Hop Question Answering on HotpotQA (EM, FBERT)
Loading...
30.81
Exact Match (EM)
CoT
8.5644
14.3397
20.115
25.8903
Apr 20, 2026
Exact Match (EM)
FBERT Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Exact Match (EM)
FBERT Score
CoT
SLM=Qwen3-235B
2026.04
30.81
69.26
Dynamic Collaboration Framework
SLM=Qwen3-8B
2026.04
29.66
69.21
Dynamic Collaboration Framework
SLM=Llama-3.1-8B
2026.04
25.54
67.06
Dynamic Collaboration Framework
SLM=Qwen3-4B
2026.04
25.51
67.3
Direct Infer
SLM=Qwen3-235B
2026.04
25
65.35
Dynamic Collaboration Framework
SLM=Llama-3.2-3B
2026.04
22.77
64.96
CoT
SLM=Qwen3-8B
2026.04
15.15
59.59
CoT
SLM=Llama-3.1-8B
2026.04
14.82
55.31
CoT
SLM=Qwen3-4B
2026.04
10.98
57.41
CoT
SLM=Llama-3.2-3B
2026.04
9.42
51.36
Feedback
Search any
task
Search any
task