SOTA Paired-prompt evaluation benchmarks and papers with code

Benchmarks

Dataset Name	SOTA Method	Metric
Winoground	LLaVA-NeXT-Vicuna-7B	Simple Accuracy58.81	2	3d ago
SugarCrepe	LLaVA-NeXT-Vicuna-7B	Simple Accuracy64.56	2	3d ago
NaturalBench	LLaVA-NeXT-Vicuna-7B	Simple Accuracy67.81	2	3d ago
MME	LLaVA-NeXT-Vicuna-7B	Simple Accuracy75.7	2	3d ago
HallusionBench	LLaVA-NeXT-Vicuna-7B	Simple Accuracy52.89	2	3d ago
BEAF (sample)	LLaVA-NeXT-Vicuna-7B	Simple Accuracy90.67	2	3d ago

Showing 6 of 6 rows