Datasets › Winoground

Winoground

Introduced by Tristan Thrush et al. in Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality7 Apr 2022 archive 2025-07-28

Winoground is a dataset for evaluating the ability of vision and language models to conduct visio-linguistic compositional reasoning. Given two images and two captions, the goal is to match them correctly -- but crucially, both captions contain a completely identical set of words, only in a different order. The dataset was carefully hand-curated by expert annotators and is labeled with a rich set of fine-grained tags to assist in analyzing model performance.

Benchmarks archive 2025-07-28

All 1 leaderboard whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Visual Reasoning Winoground GPT-4o + CA Text Score 75.5 A Cognitive Paradigm Approach to Probe the... — 114 Compare

Papers archive 2025-07-28

papers with a benchmark row: 19 · with a code link: 12 · where Syntology ran a sample: 7 (6 with a run with no instrument failure, 1 where every run was a failure of Syntology's instrument) Syntology

Show: all papers with a benchmark rowonly where code ran (7 of 19 with a benchmark row: 6 with a run with no instrument failure, 1 where every run was a failure of Syntology's instrument)

Page 1 of 1: papers 1 to 19 of 19 with a leaderboard row on this dataset's benchmarks, newest first by the archive's date (ties by slug; undated papers last).

The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset, not that list; the archive's count for this dataset is 90. The Syntology column is from Syntology's graph, stated per sample; it is not part of any archive number. A line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified” (C is a part of N, never taken away from it); the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced. When the archive marks a repository official for the paper, the cell starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record.

PaperCodeResultsDateSamples run Syntology
A Cognitive Paradigm Approach to Probe the Perception-Reasoning Interface in VLMs 0 1 23 Jan 2025 not harvested
Prompting Large Vision-Language Models for Compositional Reasoning 1 3 20 Jan 2024 official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified (2 pointer-only for licence)
CoCoT: Contrastive Chain-of-Thought Prompting for Large Multimodal Models with Multiple Image Inputs 1 13 5 Jan 2024 not harvested
Compositional Chain-of-Thought Prompting for Large Multimodal Models 1 6 27 Nov 2023 official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified (1 pointer-only for licence)
SelfEval: Leveraging the discriminative nature of generative models for evaluation 0 6 17 Nov 2023 not harvested
The Role of Chain-of-Thought in Complex Vision-Language Reasoning Task 0 2 15 Nov 2023 not harvested
MMICL: Empowering Vision-language Model with Multi-Modal In-Context Learning 2 1 14 Sep 2023 official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified (7 pointer-only for licence)
An Examination of the Compositionality of Large Generative Vision-Language Models 1 5 21 Aug 2023 not harvested
Revisiting the Role of Language Priors in Vision-Language Models 1 3 2 Jun 2023 not harvested
What You See is What You Read? Improving Text-Image Alignment Evaluation 1 8 17 May 2023 not harvested
Measuring Progress in Fine-grained Vision-and-Language Understanding 2 9 12 May 2023 official (archive's flag): 3 ran · 4 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified (1 pointer-only for licence)
Simple Token-Level Confidence Improves Caption Correctness 0 6 11 May 2023 not harvested
Incorporating Structured Representations into Pretrained Vision & Language Models Using Scene Graphs 0 12 10 May 2023 not harvested
Going Beyond Nouns With Vision & Language Models Using Synthetic Data 1 3 30 Mar 2023 not harvested
Your Diffusion Model is Secretly a Zero-Shot Classifier 4 1 28 Mar 2023 official (archive's flag): 1 ran · 2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified (1 pointer-only for licence)
Equivariant Similarity for Vision-Language Foundation Models 1 6 25 Mar 2023 official (archive's flag): 1 ran · 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample that ran constructed an object rather than computing a result
ViLEM: Visual-Language Error Modeling for Image-Text Retrieval 0 2 1 Jan 2023 not harvested
Does Structural Attention Improve Compositional Representations in Vision-Language Models? 0 4 3 Dec 2022 not harvested
Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality 2 20 7 Apr 2022 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified (3 pointer-only for licence)

Dataset loaders archive 2025-07-28

No loader listed in the archive.

Tasks archive 2025-07-28

License archive 2025-07-28

No licence recorded in the archive. Absence here is not a statement about the dataset's terms.

Modalities archive 2025-07-28

Languages archive 2025-07-28

Variants archive 2025-07-28

  • Winoground

1 variant name, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections