Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-token generation on SciQ
Loading...
40.8
Accuracy
IA2 → SFT
5.544
14.697
23.85
33.003
Sep 26, 2025
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
IA2 → SFT
Backbone=Qwen3-4B-Base...
2025.09
40.8
ICL
Backbone=Qwen3-4B-Base...
2025.09
37.5
SFT only
Backbone=Qwen3-4B-Base...
2025.09
35
w/o ICL
Backbone=Qwen3-4B-Base...
2025.09
15.4
IA2 only
Backbone=Qwen3-4B-Base...
2025.09
6.9
Feedback
Search any
task
Search any
task
