metadata
dataset_info:
- config_name: ARXIV_corpus_v1
features:
- name: paper_id
dtype: string
- name: title
dtype: string
- name: authors
dtype: string
- name: year
dtype: string
- name: venue
dtype: string
- name: doi
dtype: string
- name: dblp_key
dtype: string
- name: dblp_url
dtype: string
- name: pdf_path
dtype: string
- name: dag
dtype: string
- name: num_nodes
dtype: int64
- name: num_edges
dtype: int64
splits:
- name: gpt_5_4
num_bytes: 1721507
num_examples: 20
download_size: 640443
dataset_size: 1721507
- config_name: ARXIV_corpus_v2_top200
features:
- name: paper_id
dtype: string
- name: title
dtype: string
- name: authors
dtype: string
- name: year
dtype: string
- name: venue
dtype: string
- name: doi
dtype: string
- name: dblp_key
dtype: string
- name: dblp_url
dtype: string
- name: pdf_path
dtype: string
- name: dag
dtype: string
- name: num_nodes
dtype: int64
- name: num_edges
dtype: int64
- name: num_figures
dtype: int64
splits:
- name: gpt_5_4
num_bytes: 24636793
num_examples: 189
download_size: 8829846
dataset_size: 24636793
configs:
- config_name: ARXIV_corpus_v1
data_files:
- split: gpt_5_4
path: ARXIV_corpus_v1/gpt_5_4-*
- config_name: ARXIV_corpus_v2_top200
data_files:
- split: gpt_5_4
path: ARXIV_corpus_v2_top200/gpt_5_4-*