ArxivMIA / README.md
zhliu's picture
Update README.md
34a442c verified
|
raw
history blame
487 Bytes
metadata
license: cc-by-4.0
task_categories:
  - text-classification
language:
  - en

Dataset Card for ArxivMIA

To evaluate various pre-training data detection methods in a more challenging scenario, we introduce ArxivMIA, a new benchmark comprising abstracts from the fields of Computer Science (CS) and Mathematics (Math) sourced from Arxiv.