arXiv v1

First public version of the paper, as submitted to arXiv: arXiv:2607.25572 [cs.CR].

Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion — Cédric Bonhomme and Alexandre Dulaunoy (CIRCL)

• arxiv-v1.tar.gz — the trimmed arXiv upload package (main.tex, main.bbl, the three figure PDFs, and the deployment screenshot), extracted from this tag with git archive and verified to compile standalone.
• The compiled paper (main.pdf) and the complete trainer logs behind every table (trainer-logs/) are tracked in the repository at this tag.

Datasets and models

Artifact	DOI
CIRCL/vulnerability-attack-techniques	10.57967/hf/9621
CIRCL/vulnerability-attack-techniques-llm-scaling	10.57967/hf/9622
CIRCL/vulnerability-attack-technique-classification-roberta-base	10.57967/hf/9623
CIRCL/vulnerability-attack-technique-classification-roberta-base-llm-expanded	10.57967/hf/9624

Code: vulnerability-lookup/VulnTrain