Harmful Content Detection on Standard Harmful Content Datasets (Goal Hijacking Attack) benchmark leaderboard · SOTA2 Research

Harmful Content Detection on Standard Harmful Content Datasets (Goal Hijacking Attack)

96PhishingGAVEL

Evaluation Results

Method	Links
GAVEL2026.01		96	89	87	100	99	99	86	100	90
GPT42026.01		55	70	49	90	63	91	28	48	12