Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Single-Hop Search-augmented Question Answering on TriviaQA
Loading...
61.8
Success Rate
+ MASA
45.888
50.019
54.15
58.281
May 29, 2026
Success Rate
Updated 2d ago
Evaluation Results
Method
Method
Links
Success Rate
+ MASA
Backbone=Qwen3-14B
2026.05
61.8
+ MASA
Backbone=Qwen3-32B
2026.05
61.6
+ DS-Adapter
Backbone=Qwen3-32B
2026.05
61.5
+ Base Skill
Backbone=Qwen3-32B
2026.05
61.4
+ Base Skill
Backbone=Qwen3-14B
2026.05
60.5
No Skill
Backbone=Qwen3-14B
2026.05
60.2
+ DS-Adapter
Backbone=Qwen3-14B
2026.05
60.2
No Skill
Backbone=Qwen3-32B
2026.05
59.8
+ Base Skill
Backbone=Qwen3-8B
2026.05
58.5
+ DS-Adapter
Backbone=Qwen3-8B
2026.05
57.6
+ Base Skill
Backbone=Qwen3-4B
2026.05
57.4
+ MASA
Backbone=Qwen3-8B
2026.05
56.7
+ DS-Adapter
Backbone=Qwen3-4B
2026.05
56.5
+ MASA
Backbone=Qwen3-4B
2026.05
55.3
No Skill
Backbone=Qwen3-4B
2026.05
51
No Skill
Backbone=Qwen3-8B
2026.05
46.5
Feedback
Search any
task
Search any
task