Use cases + feedback thread (features, entity types, edge cases)
#1
by
juanmcristobal - opened
Hey folks! 👋 I’m collecting real-world feedback for attack-vector/SecureModernBERT-NER.
1) Where does it fit in your workflow?
(e.g., enrichment, indexing, triage, alert context, report drafting, etc.)
2) What are your biggest pain points?
(e.g., IOCs, actor naming, tool/malware ambiguity, CVE context, false positives/negatives, messy formatting, PDFs, etc.)
3) What would make this model “production-ready” for you?
(e.g., speed, precision/recall, stability, better docs, examples, benchmarks, eval sets, deployment recipe, integrations)
Share your use case + tricky examples
If you can, drop a small redacted/anonymized snippet that causes issues (or that it handles well).
I’ll use this thread to prioritize improvements and publish updates + docs/examples. Thanks!