List your company

Hallucination Detection on SugarCrepe 1.0 (test)

98.86Avg-M ScoreHuman
62.501671.940881.3890.8192May 20, 2026

Evaluation results

MethodLinks
HumanNote=Taken from original papers2026.0598.86999799100999999
VQAScoreModel Category=LLM-based, Backbone=CLIP-FlanT5-XXL2026.059291.695.494.597.688.393.582.9
EXPERTModel Category=LLM-based2026.0589.789.289.795.689.786.590.589.8
BEiTScoreModel Category=Encoder-based, Backbone=BEiT-L/14, Stage 1=Included2026.0588.782.886.98993.885.99488.6
BEiTScoreModel Category=Encoder-based, Backbone=BEiT-L/14, Stage 1=Excluded2026.0588.372.88484.192.776.789.683.3
FLEURModel Category=LLM-based2026.05888887.593.79187.688.786.5
LongCLIP-SModel Category=Encoder-based, Backbone=ViT-L2026.0582.386.491.586.996.476.868.270.2
PAC-S++Model Category=Encoder-based, Backbone=ViT-L2026.0582.286.994.987.497.877.367.763.7
BEiTScoreModel Category=Encoder-based, Backbone=BEiT-B/32, Stage 1=Included2026.0581.976.278.381.990.177.98980
PAC-S++Model Category=Encoder-based, Backbone=ViT-B2026.0580.580.990.68595.374.570.666.5
PAC-SModel Category=Encoder-based, Backbone=ViT-L2026.0580.48391.88696.376.368.560.8
CLIP-SModel Category=Encoder-based, Backbone=ViT-L2026.0579.78391.484.996.270.87061.6
LongCLIP-SModel Category=Encoder-based, Backbone=ViT-B2026.05797684.285.794.574.870.667.4
PAC-SModel Category=Encoder-based, Backbone=ViT-B2026.0577.275.388.38293.572.668.260.8
CLIP-SModel Category=Encoder-based, Backbone=ViT-B2026.0577.17887.382.693.669.16861.2
BEiTScoreModel Category=Encoder-based, Backbone=BEiT-B/32, Stage 1=Excluded2026.0573.662.374.471.784.965.476.169.4
GPT4o (API)Model Category=LLM-based2026.0566.760.660.166.261.37977.378.4
SPECSModel Category=Encoder-based2026.0563.982.138.470.779.757.360.158.8