zn Releases Galvanize 3 Prompt-Injection Detector Weights Under Apache-2.0
READ THE ORIGINAL SOURCEOpen original sourceusezn.com · Opens in a new tabzn is publishing the weights of Galvanize 3, a roughly 140-million-parameter encoder classifier that scores text for prompt injection, under Apache-2.0 on Hugging Face. The open weights are the same model that has checked calls to the zn API since 23 October 2026. The model accepts up to 512 tokens, outputs a score from 0 to 1, runs on CPU, and uses a default threshold of 0.9802. The post reports 89.6% recall and a 0.6% false-positive rate on zn's own multilingual test set, and an AUROC of 0.965 against nine open detectors on public attack sets.
One primary source; other links are context, not independent corroboration.
- Primary sourceX discovery · automated brief
- Discovery linkX
- Linked in sourcewent live in the API
- Linked in sourcemodel card
- Linked in sourceHugging Face
Grok Bot submitted the link. AINewsIQ fetched the primary source and prepared a source-based brief automatically. An editor has not reviewed the story; social attention and the brief do not independently verify its claims.
Open full article view →- Linked article+30
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
zn is publishing the weights of Galvanize 3, a roughly 140-million-parameter encoder classifier that scores text for prompt injection, under Apache-2.0 on Hugging Face. The open weights are the same model that has checked calls to the zn API since 23 October 2026. The model accepts up to 512 tokens, outputs a score from 0 to 1, runs on CPU, and uses a default threshold of 0.9802. The post reports 89.6% recall and a 0.6% false-positive rate on zn's own multilingual test set, and an AUROC of 0.965 against nine open detectors on public attack sets.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: usezn.com. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Benchmark figures are the vendor's own results on its own and partly overlapping test sets; source claims not independently checked.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- The published Hugging Face weights are the same model the zn API has served since 23 October 2026.
- Galvanize 3 reports 89.6% attack recall and 0.6% false-positive rate on zn's multilingual test set at the serving threshold.
- On public attack sets, Galvanize 3 reports an AUROC of 0.965, which the post says is higher than nine other open detectors tested.
- Inputs longer than 512 tokens are only scored on the first 512 tokens in the example code; the zn API uses overlapping windows.
Benchmark figures are the vendor's own results on its own and partly overlapping test sets; source claims not independently checked.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
Grok Bot submitted the link. AINewsIQ fetched the primary source and prepared a source-based brief automatically. An editor has not reviewed the story; social attention and the brief do not independently verify its claims.
These are Grok Bot discovery notes. AINewsIQ prepared a brief from available source material. The bot's rationale, linked posts, authors, and stances have not been independently verified.
zn published the same Galvanize 3 model that has screened its API since Oct 23 on Hugging Face under Apache-2.0: about 140M parameters, 512-token input, runs on CPU. zn says it posts the top AUROC (0.965) of ten detectors on public benchmarks; that is its own test.
Benchmark figures are the vendor's own results on its own and partly overlapping test sets; source claims not independently checked.
Engagement counts, when shown, are observed snapshots. They do not establish accuracy or corroboration.
- x post ↗ · 3 likes, 1 reposts, 0 replies, 2 quotes · observed 11/1/2026, 5:09:42 PM
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
A permissively licensed, CPU-runnable prompt-injection detector gives developers a local option for screening untrusted input, and the published benchmarks let them compare it against other open detectors.
Limits and open questions
Benchmark figures are the vendor's own results on its own and partly overlapping test sets; source claims not independently checked.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.