Dev trains own PII detector on Nemotron, says off-the-shelf models aren't good enough
HankYeomans · x · 2026-10-16
Developer Hank Yeomans shares how he built his own PII detector: he started with ONNX and GLiNER plus off-the-shelf PII datasets, then moved to Nemotron after joining NVIDIA Inception and discovering the Nemotron PII dataset. Believing off-the-shelf options fall short and the good stuff is locked behind commercial products, he is now fine-tuning Nemotron as a base with his own corpus — still refining and open to better alternatives.
Related event: Developer trains own PII detector, unhappy with existing tools(3 posts)→
More from Models
- Muse Wonder scores 14.9% on WeirdML v3, strongest non-GPT/Claude model yet — scaling01 · 2026-10-17
- Zed CEO: OpenAI's Dots is a low-effort model that abandons tasks midway — zeeg · 2026-10-17
- Agent Arena: OpenAI and Anthropic Lead on User Enjoyment as Labs Converge on Reliability — arena · 2026-10-17
- Google AI Studio Build gains Max and Plan modes as Gemini 4 Argon surfaces in UI — testingcatalog · 2026-10-17
- NEURA launches Neuron AI, a multimodal PLA model giving robots touch, force feedback and memory — CyberRobooo · 2026-10-17
- Homemade conflicting-tests experiment: 4 models flagged it, Gemini 3.1 Pro faked green builds 20/20 — Nir777 · 2026-10-17