---
language:
- en
- zh
license: other
size_categories:
- 1K<n<10K
task_categories:
- text-generation
license_name: apache-2.0-and-upstream-licenses
license_link: LICENSE.md
tags:
- verifier-generation
- code
- evaluation
- benchmark
configs:
- config_name: tasks
  data_files:
  - split: full
    path: data/tasks.parquet
  default: true
- config_name: candidates
  data_files:
  - split: full
    path: data/candidates.parquet
---

# VerifierBench

[**Paper**](https://huggingface.co/papers/2610.28679) · [**Project page**](https://verifierbench.github.io/) · [**Code**](https://github.com/DwanZhang-AI/verifier-suite)

VerifierBench provides the paper’s frozen 222 tasks and 4,993 constructed candidates with PASS/FAIL labels. AppWorld content requires [local reconstruction](USAGE.md); source terms are in [LICENSE.md](LICENSE.md).

```python
from datasets import load_dataset

tasks = load_dataset("Catherine1213/VerifierBench", "tasks", split="full")
candidates = load_dataset("Catherine1213/VerifierBench", "candidates", split="full")
```