AI RESEARCH
Pretraining Data Exposure in Large Language Models: A Survey of Membership Inference, Data Contamination, and Security Implications
arXiv CS.AI
•
ArXi:2605.26133v1 Announce Type: cross Large Language Models (LLMs) have become the predominant paradigm in NLP, advancing both research and industry. As model sizes and pre