Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Preference Alignment on Argilla-7k (test)
Loading...
9.23
LC Win Rate
MixDPO
2.6468
4.3559
6.065
7.7741
Feb 1, 2026
LC Win Rate
Win Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
LC Win Rate
Win Rate
MixDPO
Base Model=LLaMA-3-8B
2026.02
9.23
20.62
SimPO
Base Model=LLaMA-3-8B
2026.02
7.07
5.84
BASE SFT
Base Model=LLaMA-3-8B
2026.02
3.73
10.19
SELECTIVEDPO
Base Model=LLaMA-3-8B
2026.02
3.59
5.22
VANILLA DPO
Base Model=LLaMA-3-8B
2026.02
2.9
12.17
Feedback
Search any
task
Search any
task