Some data files seem to contain viruses?
According to the ClamAV - https://www.clamav.net/ - open source virus scanner used by Hugging Face some files contain viruses or similar.
Dos.Trojan.DeltreeY-28
See for example
https://huggingface.co/datasets/GPT-NL/GPT-NL_Public_Corpus/blob/main/cc_github_open_source/cc_github_open_source_0643.parquet
You are probably already aware.
Do you intent to filter those in an update or shall we remove them when re-using the dataset.
Where those also in when training GPT-NL or where those excluded?
cc_github_open_source_0643.parquet : Dos.Trojan.DeltreeY-28
cc_github_open_source_0796.parquet : Dos.Trojan.DeltreeY-32
cc_github_open_source_0934.parquet : Dos.Trojan.Deltree-32
cc_github_open_source_1009.parquet : Win.Exploit.Alpha_Upper-1
cc_github_open_source_1029.parquet : Html.Exploit.USBank-1
cc_github_open_source_1319.parquet : Legacy.Trojan.Agent-1388612
cc_github_open_source_1764.parquet : Win.Exploit.Alpha_Upper-1
cc_github_open_source_2324.parquet : Win.Exploit.Alpha_Upper-1
Hello! Thanks for checking out the corpus. We will keep this in mind when updating the corpus. We did use these files as well for training the LLM. As code generation is not one of the main tasks for GPT-NL, we don't think this will be a vulnerability, but we are red-teaming for these type of issues.