openai/gpt-oss-120b pass
No major issues. Minor: a small glitch-token surface (non-English text only).
chat template: present · view on Hugging Face ↗
Scan coverageStatic battery2026-08-22Weights battery2026-08-20Behavioral batterynot rundetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-22 |
| Weights battery | Weights forensics: no GPU, no download | complete 2026-08-20201,088-token embedding scanned · 936 undertrained |
| Behavioral battery | Live-inference differentials | not run |
Ingot runs three batteries against a model. What each one checks →
Findings
Scanned 2026-08-22 · published from a community scan.
low Undertrained tokens in vocabulary (non-ASCII tail)
Embedding-norm scan flagged 936 undertrained tokens (norm < 0.3× the vocabulary median), but 0 decode to plain-ASCII strings, so exposure in English-language pipelines is limited. Behavioral confirmation requires the behavioral battery.
How to fixruntime guardweight-level
Keep the affected token strings out of the model's input — the scan-derived runtime guard carries this model's exact blocklist.
- Fetch this model's guard artifact (`/api/v1/guard/<owner>/<model>`): the confirmed corrupting tokens and the low-norm candidate list, derived from the published scan.
- Screen inbound text with it (the `@ingotai/guard` package is a reference implementation) and route flagged records to a different model or human review — verbatim-copy tasks on flagged strings are the failure mode.
- The underlying cause is undertrained embeddings in the weights; a true fix is weight-level (continued pretraining on the affected tokens) — that is not a patch, it's a training job.
info Glitch-token echo probe: no candidate surface
Embedding-norm scan found no undertrained ASCII tokens (threshold 28.2629 vs vocab median 94.2097), so the glitch-token data-corruption class has nothing to trigger on. Echo probe not applicable.
Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.
Check every checkpoint before it ships
Use the web app, API, CLI, or CI gate to scan candidate checkpoints and catch model drift before deployment. Public-model scans publish to the open database; paid plans add the volume needed for continuous checks.
Fingerprint
The durable profile of this model: measured weights-and-metadata facts, rebuilt on every scan and battery run. Updated 2026-08-20.
| architecture | gpt_oss · 36 layers · 2880-dim |
| parameters | 116829.2M |
| vocabulary | 201,088 tokens |
| license | apache-2.0 |
| serialization | safetensors pickle |
| chat template | none |
| glitch-token surface | 936 undertrained candidates, 0 plain-ASCII |
Full measured fingerprint
| architectures | GptOssForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 37 — pickle: metal/model.bin |
| revision | b5c939de8f75 |
| HF snapshot | 4.7M downloads · 5.1k likes · updated 2025-08-26 · captured 2026-08-20 |
| embedding tensor | model.embed_tokens.weight · BF16 · 201,088×2880 |
| embedding norms | median 94.2097 · mean 94.4033 |
| lineage check | no claimed base model |
Battery runs (4)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| gpu | complete | 2026-08-22 04:16 | 80s | 1 |
| weights | complete | 2026-08-20 18:02 | 28s | 1 |
| weights | complete | 2026-08-20 08:10 | 26s | 1 |
| weights | complete | 2026-08-20 06:35 | 38s | 1 |
gpu run 2026-08-22 — measurements
| probes run | glitch |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/openai/gpt-oss-120b)