Visual Reasoning on Winoground

114개 결과 · ⬇ CSV · JSON

Text Score

Rank	Model	Text Score	Image Score	Group Score	Paper	Code	Year
1	GPT-4o + CA	75.5	58.5	52	A Cognitive Paradigm Approach to Probe the Perception-Reasoning Interface in VLMs		2025
2	GPT-4V (CoT, pick b/w two options)	75.25	68.75	58.75	The Role of Chain-of-Thought in Complex Vision-Language Reasoning Task		2023
3	GPT-4V (pick b/w two options)	69.25	46.25	39.25	The Role of Chain-of-Thought in Complex Vision-Language Reasoning Task		2023
4	MMICL + CoCoT	64.25	52.5	50.75	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
5	GPT-4V + CoCoT	58.5	49.5	44.5	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
6	OpenFlamingo + CoCoT	58.25	55.25	41.5	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
7	GPT-4V	54.5	42.5	37.75	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
8	FIBER (EqSim)	51.5	32.00	27.5	Equivariant Similarity for Vision-Language Foundation Models	wangt-cn/eqben	2023
9	FIBER (finetuned, Flickr30k)	51.25	26.50	23.00	Equivariant Similarity for Vision-Language Foundation Models	wangt-cn/eqben	2023
10	MMICL + CCoT	51	48	47.5	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
11	OpenFlamingo + DDCoT	47.5	47.25	39	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
12	VQ2	47	42.2	30.5	What You See is What You Read? Improving Text-Image Alignment Evaluation	yonatanbitton/wysiwyr	2023
13	MMICL + DDCoT	46.75	45	36.75	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
14	X-VLM 16M	46.7	24.5	21.2	Measuring Progress in Fine-grained Vision-and-Language Understanding	e-bug/fine-grained-evals · e-bug/weak-relation-vlm	2023
15	PaLI (ft SNLI-VE + Synthetic Data)	46.5	38	28.75	What You See is What You Read? Improving Text-Image Alignment Evaluation	yonatanbitton/wysiwyr	2023
16	FIBER	46.25	25.75	22.25	Equivariant Similarity for Vision-Language Foundation Models	wangt-cn/eqben	2023
17	MMICL (FLAN-T5-XXL)	45.50	44.99	43.00	MMICL: Empowering Vision-language Model with Multi-Modal In-Context Learning	haozhezhao/mic · pkunlp-icler/mic	2023
18	PaLI (ft SNLI-VE)	45.00	41.50	28.70	What You See is What You Read? Improving Text-Image Alignment Evaluation	yonatanbitton/wysiwyr	2023
18	Gemini + DDCoT	45	25	23.75	CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs	vista-h/gpt-4v_social_media	2024
18	METER (EqSim)	45.0	22.75	18.75	Equivariant Similarity for Vision-Language Foundation Models	wangt-cn/eqben	2023