Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Boolean Question Answering on BoolQ (acc_norm)
Loading...
85.3
Acc (Normalized)
Base
-18.908
8.146
35.2
62.254
Apr 6, 2026
Acc (Normalized)
Updated 11d ago
Evaluation Results
Method
Method
Links
Acc (Normalized)
Base
Model=Llama-70B
2026.04
85.3
Base
Model=Mistral-7B
2026.04
84.3
Base
Model=Qwen2.5-7B
2026.04
84.3
Base
Model=Llama-13B
2026.04
82.5
Base
Model=GPT-2 (774M)
2026.04
60.5
GAIN
Model=Qwen2.5-7B
2026.04
-0.8
GAIN
Model=Mistral-7B
2026.04
-1.2
GAIN
Model=Llama-13B
2026.04
-1.4
GAIN
Model=GPT-2 (774M)
2026.04
-2.1
LoRA
Model=Llama-13B
2026.04
-2.3
GAIN
Model=Llama-70B
2026.04
-2.5
LoRA
Model=GPT-2 (774M)
2026.04
-7.8
LoRA
Model=Mistral-7B
2026.04
-9.2
LoRA
Model=Llama-70B
2026.04
-14.6
LoRA
Model=Qwen2.5-7B
2026.04
-14.9
Feedback
Search any
task
Search any
task
