---
pretty_name: Generative Embedding Benchmark
license: apache-2.0
task_categories:
- visual-question-answering
---

# Generative Embedding Benchmark

<div align="center">
  <a href="https://arxiv.org/abs/2608.06972"><img src="https://img.shields.io/badge/arXiv-2608.06972-b31b1b.svg" alt="arXiv"></a>
  <a href="https://github.com/LimitedMouse/Generative-Embedding-Benchmark"><img src="https://img.shields.io/badge/GitHub-Code-lightgrey" alt="GitHub Code"></a>
  <a href="https://huggingface.co/LimitedMouse/Generative-Embedding-Benchmark-Checkpoints"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Checkpoints-yellow" alt="Hugging Face Checkpoints"></a>
</div>

This repository publishes the fixed membership and provenance metadata for the
Generative Embedding Benchmark (GEB). GEB contains 1,800 development questions
and 900 held-out test questions spanning natural images, scene text, and visual
documents.

GEB evaluates how much answer-relevant visual information a dense embedding
makes accessible to a generative decoder. The
[GitHub repository](https://github.com/LimitedMouse/Generative-Embedding-Benchmark)
is the main entry point for installation, training, and evaluation. Official
Qwen3-VL-Embedding-2B/8B decoder weights are provided in the
[checkpoint repository](https://huggingface.co/LimitedMouse/Generative-Embedding-Benchmark-Checkpoints).

This dataset repository does not redistribute source images, questions,
answers, or decoder training data. GEB task definitions load the eight original
public evaluation datasets through `lmms-eval` and filter them with `dev.json`
or `test.json`. The source datasets remain subject to their own licenses and
access requirements.

## Files

- `dev.json`, `test.json`: compact source-row membership used at runtime.
- `manifests/dev/`, `manifests/test/`: categories, provenance identifiers,
  image hashes, and image-group identifiers used for auditing.
- `sources.json`: pinned upstream dataset revisions.
- `summary.json`: sampling quotas and aggregate attrition statistics.
- `leakage_audit.json`: development/test image-leakage audit.

## Citation

```bibtex
@misc{li2026generativeembeddingbenchmark,
  title={Generative Embedding Benchmark: How Much Information Survives in a Dense Embedding?},
  author={Yun Li and Biao Yang and Peixi Wu and Yunhao Zhou and Mingzhou Jiang and Wei Yuan and Fan Yang and Wenwu Ou},
  year={2026},
  eprint={2608.06972},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2608.06972},
}
```

The Apache-2.0 license applies only to GEB-authored manifests and metadata. It
does not alter the terms of the upstream datasets.