"Evading Data Contamination Detection for Language Models is (too) Easy."

Jasper Dekoninck et al. (2024)

Details and statistics

DOI: 10.48550/ARXIV.2402.02823

access: open

type: Informal or Other Publication

metadata version: 2024-02-12

a service of  Schloss Dagstuhl - Leibniz Center for Informatics