File size: 2,735 Bytes
930a238 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 | ---
license: cc-by-4.0
language:
- en
task_categories:
- question-answering
pretty_name: ElephantBench
size_categories:
- 1K<n<10K
configs:
- config_name: default
data_files:
- split: test
path: data/elephantbench.jsonl
---
# ElephantBench
<p align="center">
<a href="https://github.com/Tencent/ElephantBench">
<img
src="https://img.shields.io/badge/ElephantBench-GitHub-blue?logo=github"
alt="GitHub Repo"
/>
</a>
<a href="https://tencent.github.io/ElephantBench/">
<img
src="https://img.shields.io/badge/ElephantBench-Leaderboard-ff725e?logo=githubpages&logoColor=white"
alt="ElephantBench Leaderboard"
/>
</a>
<a href="https://arxiv.org/abs/2608.28478">
<img
src="https://img.shields.io/badge/ElephantBench-Paper-red?logo=arxiv&logoColor=red"
alt="Paper"
/>
</a>
</p>
ElephantBench is a closed-book knowledge probe for evaluating whether a language model
remembers long-tail facts and recalls the different verified accounts associated with them.
The release contains 1,094 English questions.
Evaluation code, prompts, construction utilities, and full documentation are available in
the [ElephantBench GitHub repository](https://github.com/Tencent/ElephantBench).
## Load the dataset
```python
from datasets import load_dataset
dataset = load_dataset("Tencent/ElephantBench", split="test")
```
## Record format
```json
{
"benchmark_id": "4382fd6d-ec5d-5ec8-a3f2-8abc706fa010",
"item_group_id": "c9ef1c87-3e50-5190-8e8e-37a82ec25634",
"eval": {
"question": "What birth date was reported for Mother Teresa?",
"gold_answers": [
{"value": "August 26, 1910"},
{"value": "August 27, 1910"}
],
"preferred_answer": "Reports cite August 26 and August 27, 1910."
}
}
```
The target model receives only `eval.question`. Gold answers are supplied to the judge after
generation.
## Evaluation
- **C (complete):** all verified answers are covered without a material contradiction.
- **P (partial):** at least one, but not all, verified answers are covered.
- **F (failed):** no verified answer is covered, a material contradiction is present, or the
generation/judging request failed.
- **K (conditional completeness):** `C / (C + P)`.
## License
[cc-by-4.0](https://huggingface.co/datasets/tencent/ElephantBench/blob/main/LICENSE).
## Citation
```
@article{pan2026elephantbench,
title={Blind Men and the Elephant: Probing the Epistemic Myopia of LLMs under Long-Tail Divergent Knowledge},
author={Pan, Zhuoshi and Lu, Junru and Qian, Yan and Zhao, H. Vicky and Yin, Di and Sun, Xing},
journal={arXiv preprint arXiv:2608.28478},
year={2026}
}
``` |