○MODEL CARD · ENTRY 03 — NO WITNESS · NO ENTRY
GPT-2 · the abstain control
The control that proves the catches mean something. GPT-2 has no safety behavior to strip, so there is nothing to detect — and Protora returns no finding. An incumbent raw-embedding baseline false-fires here with nothing to stand behind it; Protora abstains. A tool that never abstains cannot be trusted on the runs where it does fire.
Nothing to disclose, nothing to detect — the abstain is the entry.
No safety behavior to strip — so no finding is manufactured. The honest blank is what makes every catch on the record mean something.
No refusal rate to read — the subject is a base model.
○THE FULL RECORD
The record behind the read.
Everything the read is built on — the provenance, the claims, the evidence, the signed attestation. Depth, not a second telling: pick a tab.
○ HONEST SCOPE
What this card does not claim.
The misses are part of the card, not an appendix. Every catch above sits next to what could not be proven — and it never renders folded away.
- No safety behavior present — nothing to detect, so nothing is reported.
- This is the whole entry: the honest gap, shown as loudly as the catches.
COVERAGE — the abstain is the product: no witness, no finding. The honesty anchor for every catch on the record.
Excise starts from a proven finding. No finding, no patch. Fail closed.
On-ramps: this catch on the chronological record, the deepest read, signed OPEN NOW, the full atlas, how this read scores against the open-source scanners, and why it’s called Protora.
Run it on the model
you can’t vouch for.
This card started as a model someone downloaded and trusted. Bring yours — Protora will tell you what it did, prove it with a replayable witness, and tell you plainly what it couldn’t prove.
Every engagement is scoped per case and quoted per offer — tell us what you need, and we’ll scope it and send you an offer.