You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If you find this work relevant or helpful to your work, please kindly cite us:
@inproceedings{
processbench,
title={ProcessBench: Identifying Process Errors in Mathematical Reasoning},
author={
Chujie Zheng and Zhenru Zhang and Beichen Zhang and Runji Lin and Keming Lu and
Bowen Yu and Dayiheng Liu and Jingren Zhou and Junyang Lin
},
booktitle={The 63rd Annual Meeting of the Association for Computational Linguistics
},
year={2025}
}
Data Usage
You can use the following code to preview the ProcessBench data:
importjsonfromdatasetsimportload_datasetdataset=load_dataset('Qwen/ProcessBench', split='gsm8k')
print(json.dumps(dataset[0], indent=2))
# Expected output:"""{ "id": "gsm8k-0", "generator": "Qwen2-7B-Instruct", "problem": "Sue lives in a fun neighborhood...", "steps": [ "To find out how many more pink plastic flamingos were out than...", ... ], "final_answer_correct": false, "label": 1}"""
Evaluation
You can refer to the code folder for the evaluation code and the prompt templates we use in this work
About
Official repository for ACL 2025 paper "ProcessBench: Identifying Process Errors in Mathematical Reasoning"