List your company

Text Classification on MIND

77.43AccuracyWARDEN
74.757275.451176.14576.8389Mar 3, 2024Jul 9, 2024Nov 15, 2024Mar 23, 2025Jul 30, 2025Dec 5, 2025Apr 13, 2026

Evaluation results

MethodLinks
WARDENAttack=Dimension-shift Attack2026.0477.43—10
WARDENAttack=No Attack2026.0477.39—10
WARDENAttack=Dim2025.1277.33——
OriginalAttack=None2026.0477.33—0.99
SemMarkAttack=Dim2025.1277.29——
EmbMarkerAttack=No Attack2025.1277.26——
EspeWAttack=Dim2025.1277.25——
SemMarkAttack=No Attack2025.1277.25——
WARDENAttack=Dimension-reduction Attack2026.0477.25—10
EmbMarkerAttack=Dim2025.1277.24——
EspeWAttack=Dimension-shift Attack2026.0477.24—10
OriginalDescription=Benign victim model2024.0377.2251.37—
EmbMarkerAttack=Detect2025.1277.22——
GeoMarkAttack=Dimension-shift Attack2026.0477.22—10
OriginalAttack=None2025.1277.21——
WARDENAttack=No Attack2025.1277.21——
GeoMarkAttack=No Attack2026.0477.21—10
WARDENAttack=Detect2025.1277.2——
EmbMarkerDescription=Watermarked model2024.0377.1951.4—
EspeWAttack=Detect2025.1277.18——
EspeWAttack=No Attack2025.1277.17——
EspeWAttack=No Attack2026.0477.17—10
SemMarkAttack=Detect2025.1277.09——
GeoMarkAttack=Dimension-reduction Attack2026.0477.06—10
EspeWAttack=Dimension-reduction Attack2026.0477—10
WARDENAttack=Paraphrasing Attack (GPT-4o-mini)2026.0476.93—10
WETAttack=Dimension-shift Attack2026.0476.91—0.72
WETAttack=No Attack2025.1276.89——
EspeWAttack=Paraphrasing Attack (GPT-4o-mini)2026.0476.88—0.002
GeoMarkAttack=Paraphrasing Attack (GPT-4o-mini)2026.0476.85—10
WETAttack=Detect2025.1276.83——
WETAttack=No Attack2026.0476.83—10
WETAttack=Paraphrasing Attack (GPT-4o-mini)2026.0476.68—10
WARDENAttack=Paraphrasing Attack (NLLB)2026.0476.58—10
EspeWAttack=Paraphrasing Attack (NLLB)2026.0476.57—0.002
GeoMarkAttack=Paraphrasing Attack (NLLB)2026.0476.49—10
WETAttack=Paraphrasing Attack (NLLB)2026.0476.25—10
WETAttack=Dim2025.1275.94——
WARDENAttack=CSE Attack2026.0475.69—0.05
EspeWAttack=CSE Attack2026.0475.59—10
CSEDescription=CSE performed on EmbMarker embeddings2024.0375.5150.35—
SemMarkAttack=CSE2025.1275.49——
EmbMarkerAttack=CSE2025.1275.48——
EspeWAttack=CSE2025.1275.48——
WETAttack=CSE2025.1275.37——
WARDENAttack=CSE2025.1275.28——
WETAttack=CSE Attack2026.0475.12—10
GeoMarkAttack=CSE Attack2026.0474.86—10