List your company

Workflow Steering Defense on MisinfoTask

75.92TASRG-Safeguard
17.180832.430447.6862.9296May 12, 2026

Evaluation results

MethodLinks
G-SafeguardModel=Gemini-2.5-Flash2026.0575.9273.15
ARGUSModel=Gemini-2.5-Flash2026.0564.8173.15
w/o DefenseModel=GPT-4o-mini2026.0563.8965.74
w/o DefenseModel=Gemini-2.5-Flash2026.0563.8963.89
ARGUSModel=GPT-4o-mini2026.0560.1961.11
G-SafeguardModel=GPT-4o-mini2026.0555.5663.89
w/o DefenseModel=DeepSeek-V32026.0551.8556.48
ARGUSModel=Qwen-3.5-Flash2026.055053.7
G-SafeguardModel=DeepSeek-V32026.0548.1553.7
G-SafeguardModel=Qwen-3.5-Flash2026.0547.2252.78
w/o DefenseModel=Qwen-3.5-Flash2026.0546.2951.85
ARGUSModel=DeepSeek-V32026.0541.6749.07
FLOWGUARDModel=Gemini-2.5-Flash2026.0535.1931.48
FLOWGUARDModel=GPT-4o-mini2026.0533.3335.19
FLOWGUARDModel=Qwen-3.5-Flash2026.0527.7838.89
FLOWGUARDModel=DeepSeek-V32026.0519.4422.22