List your company

Safety and Helpfulness Evaluation on WildTeaming

0.3Harm RatePPO SafeSearch w/o qr. & hf.
-3.2420.65544.5568.445Oct 19, 2025

Evaluation results

MethodLinks
PPO SafeSearch w/o qr. & hf.Backbone=Qwen-2.5-7B-Instruct2025.100.31.38
GRPO SafeSearchBackbone=Qwen-2.5-7B-Instruct2025.100.62.23
PPO SafeSearchBackbone=Qwen-2.5-7B-Instruct2025.101.92.16
PPO SafeSearch w/o hf.Backbone=Qwen-2.5-7B-Instruct2025.101.91.45
PPO SafeSearch w/o qr.Backbone=Qwen-2.5-7B-Instruct2025.103.62.25
PPO SafeSearchBackbone=Qwen-2.5-3B-Instruct2025.1048.92.51
GRPO Utility-Only Ft.Backbone=Qwen-2.5-7B-Instruct2025.1078.82.53
BaseBackbone=Qwen-2.5-7B-Instruct2025.1083.92.79
Search BlockBackbone=Qwen-2.5-7B-Instruct2025.1084.72.66
Document FilterBackbone=Qwen-2.5-7B-Instruct2025.1085.72.82
PPO Utility-Only Ft.Backbone=Qwen-2.5-7B-Instruct2025.1087.32.88
BaseBackbone=Qwen-2.5-3B-Instruct2025.1087.52.9
Document FilterBackbone=Qwen-2.5-3B-Instruct2025.10882.89
PPO Utility-Only Ft.Backbone=Qwen-2.5-3B-Instruct2025.1088.62.87
Search BlockBackbone=Qwen-2.5-3B-Instruct2025.1088.82.97