mdocekal's picture
Upload dataset
b5fbbe2 verified
---
dataset_info:
features:
- name: id
dtype: string
- name: sentence
dtype: string
- name: context
dtype: string
- name: label
dtype:
class_label:
names:
'0': none
'1': used
'2': created
'3': mention
splits:
- name: train
num_bytes: 1417050
num_examples: 2627
- name: validation
num_bytes: 178889
num_examples: 328
- name: test
num_bytes: 177860
num_examples: 328
download_size: 1050344
dataset_size: 1773799
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
---
# Software intent Softcite SoMeSci CZI dataset
This is a dataset for citation intent classification in software papers. This is just Hugging Face dataset wrapper around the original dataset.
The original dataset is available at https://docs.google.com/spreadsheets/d/1yY1vB_v-1FjVV-VQ8rye84wB8dN5kGh8da4fom8KQwk/edit?gid=1956523166#gid=1956523166 .
The GitHub repository with the original GitHub repository is available at https://github.com/karacolada/SoftwareImpactHackathon2023_SoftwareCitationIntent.
The train, validation, and test splits are not provided in the original dataset, thus we made those splits ourselves.
## Fields description
Both configurations contain the following fields:
- `id`: The unique identifier of the example.
- `orig_id`: The original identifier of the example.
- `sentence`: Sentence containing the citation.
- `context`: Broader context of the sentence. (<leading sentence><citing sentence>< trailing sentence>)
- `label`: Citation intent label.
- `0`: No citation intent.
- `1`: used
- `2`: created
- `3`: mention
## Citation
```bibtex
@inproceedings{Istrate2024ScientificSC,
title={Scientific Software Citation Intent Classification Using Large Language Models},
author={Ana-Maria Istrate and Joshua Fisher and Xinyu Yang and Kara Moraw and Kai Li and Donghui Li and Martin Klein},
booktitle={NSLP},
year={2024},
url={https://api.semanticscholar.org/CorpusID:271971376}
}
```