This dataset is an oversampled version of s2e-lab/RegexEval (Re(gEx|DoS)Eval).
The upstream dataset contains 762 curated regex prompts from real users, with match/non-match test strings. This repo repeats those rows with replacement to provide 2000 prompts for seeding CWE-1333 secure-coding task generation.
Source: s2e-lab/RegexEval
Method: Random oversampling with replacement (seed=42)
Original unique rows: 762
Oversampled… See the full description on the dataset page:
https://huggingface.co/datasets/AetherPrior/RegexEval.
