Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Sycophancy Evaluation on SycophancyEval
Loading...
54.2
Sycophancy Rate
Lag-DPO
52.464
64.182
75.9
87.618
Apr 1, 2026
Sycophancy Rate
Consistency
Updated 17d ago
Evaluation Results
Method
Method
Links
Sycophancy Rate
Consistency
Lag-DPO
Base Model=Qwen3-4B-In...
2026.04
54.2
45.8
Vanilla
Base Model=Qwen3-4B-In...
2026.04
87.2
12.9
SFT-Combined
Base Model=Qwen3-4B-In...
2026.04
87.6
12.5
SafeRLHF
Base Model=Qwen3-4B-In...
2026.04
87.8
12.2
Multi-Neg DPO
Base Model=Qwen3-4B-In...
2026.04
89.4
10.6
PCGrad-DPO
Base Model=Qwen3-4B-In...
2026.04
89.6
10.4
MODPO
Base Model=Qwen3-4B-In...
2026.04
96.2
3.8
SACPO
Base Model=Qwen3-4B-In...
2026.04
97.4
2.6
SFT-Anchor
Base Model=Qwen3-4B-In...
2026.04
97.6
2.4
Feedback
Search any
task
Search any
task