File size: 476 Bytes
9b31838 34a442c 010331e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 |
---
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.
- **Repository:** https://github.com/zhliu0106/probing-lm-data
- **Paper:** Probing Language Models for Pre-training Data Detection
|