Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Classification on WildGuard (test)
Loading...
88.5
F1 Score
TPC
58.236
66.093
73.95
81.807
Sep 30, 2025
Nov 9, 2025
Dec 20, 2025
Jan 30, 2026
Mar 11, 2026
Apr 21, 2026
Jun 1, 2026
F1 Score
Updated 1d ago
Evaluation Results
Method
Method
Links
F1 Score
TPC
Params=1,381,889
2025.09
88.5
MLP probe
Params=1,382,147
2025.09
88.2
Bilinear probe
Params=344,129
2025.09
87.82
EE-MLP
Params=1,381,918
2025.09
87.76
Linear probe
Params=5,377
2025.09
86.86
gpt-4o-mini
Params=≈ 8,000,000,000*
2025.09
86.63
claude-3-haiku
Params=≈ 8,000,000,000*
2025.09
83.24
o3-mini
Params=unknown
2025.09
82.03
SentGuard
backbone=Qwen3-4B-Inst...
2026.06
80
Qwen3Guard-8B
2026.06
77.2
Llama-Guard-3-8B
Params=8,030,261,248
2025.09
76.85
WildGuard-7B
2026.06
75.3
GPT-5.5
protocol=zero-shot
2026.06
74.4
LlamaGuard3-8B
2026.06
70.7
shieldgemma-2b
Threshold=optimal, Par...
2025.09
69.9
LlamaGuard4-12B
2026.06
66.7
Gemini-3.5-Flash
protocol=zero-shot
2026.06
59.4
Feedback
Search any
task
Search any
task