Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
LLM Alignment on HH-RLHF 100K samples (test)
Loading...
82.3
Helpfulness Score
Hard-Pair-GRPO
78.036
79.143
80.25
81.357
May 7, 2026
Helpfulness Score
Harmlessness Score
Updated 26d ago
Evaluation Results
Method
Method
Links
Helpfulness Score
Harmlessness Score
Hard-Pair-GRPO
Base Model=LLaMA-2-7B-...
2026.05
82.3
85.7
ORPO
Base Model=LLaMA-2-7B-...
2026.05
80.5
83.2
DPO
Base Model=LLaMA-2-7B-...
2026.05
80.1
82.8
Soft-Pair-GRPO
Base Model=LLaMA-2-7B-...
2026.05
79.5
82.1
Standard GRPO
Base Model=LLaMA-2-7B-...
2026.05
78.2
81.5
Feedback
Search any
task
Search any
task