zhliu/ArxivMIA
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
0317
Update README.md
Update README.md
Update README.md
Delete arxiv_mia_test.jsonl
Delete arxiv_mia_dev.jsonl
Delete arxiv_mia.jsonl
Upload 3 files
Update README.md
Update README.md
Rename arxiv_mia_all.jsonl to arxiv_mia.jsonl
Rename arxiv_mia.jsonl to arxiv_mia_all.jsonl
Upload 3 files
Update README.md
initial commit
