Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multiple-choice factual recall on POLYFACT Low
Loading...
56.93
Accuracy
GRPO
47.154
49.692
52.23
54.768
Jun 4, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
GRPO
Backbone=Qwen-2.5-7B
2026.06
56.93
GRPO
Backbone=OLMo-2-1124-7B
2026.06
54.48
CPT + GRPO
Backbone=OLMo-2-1124-7B
2026.06
54.41
Baseline
Backbone=Qwen-2.5-7B
2026.06
52.26
CPT + SFT
Backbone=OLMo-2-1124-7B
2026.06
52.02
CPT
Backbone=OLMo-2-1124-7B
2026.06
51.82
Baseline
Backbone=OLMo-2-1124-7B
2026.06
51.8
CPT
Backbone=Qwen-2.5-7B
2026.06
51.09
SFT
Backbone=OLMo-2-1124-7B
2026.06
50.04
CPT + GRPO
Backbone=Qwen-2.5-7B
2026.06
49.33
SFT
Backbone=Qwen-2.5-7B
2026.06
48.88
CPT + SFT
Backbone=Qwen-2.5-7B
2026.06
47.53
Feedback
Search any
task
Search any
task