microsoft/phi-4 warn
chat template: present · view on Hugging Face ↗
Scan coverage
Ingot runs three batteries against a model. What each one checks →
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-22 |
| Weights battery | Weights forensics — no GPU, no download | complete 2026-08-22100,352-token embedding scanned · 110 undertrained |
| Behavioral battery | Live-inference differentials | not run |
Findings
Scanned 2026-08-22 · published from a community scan.
medium Undertrained (glitch) token surface in vocabulary
Embedding-norm scan flagged 110 undertrained tokens (norm < 0.3× the vocabulary median of 1.146), including 85 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "useRalative", "useRal", "ForCanBeConvertedToF", "webElementX", "ForCanBeConverted", "PostalCodesNL", "Hexatrigesimal", "externalActionCode". In models where this class was tested behaviorally, such tokens silently rewrote user input into confident, schema-valid, wrong output. These are candidates from the weights alone; 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.
Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.
Fingerprint
The durable profile of this model: measured weights-and-metadata facts, rebuilt on every scan and battery run. Updated 2026-08-22.
| architecture | phi3 · 40 layers · 5120-dim |
| parameters | 14659.5M |
| vocabulary | 100,352 tokens |
| license | mit |
| serialization | safetensors |
| chat template | present · sha256:fe4120415a3858b8 |
| glitch-token surface | 110 undertrained candidates, 85 plain-ASCII |
Full measured fingerprint
| architectures | Phi3ForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 21 |
| revision | 2db69c1c3e91 |
| HF snapshot | 695.6k downloads · 2.3k likes · updated 2026-07-14 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · BF16 · 100,352×5120 |
| embedding norms | median 1.1463 · mean 1.1121 |
| lineage check | no claimed base model |
| glitch-token samples | "useRalative", "useRal", "ForCanBeConvertedToF", "webElementX", "ForCanBeConverted", "PostalCodesNL", "Hexatrigesimal", "externalActionCode", "davidjl", "artisanlib", ".bunifuFlatButton", "thuisontvangst" |
Battery runs
The run trace behind the findings above: every deep-battery job for this model, with what each run measured or why it failed. Findings are only as good as the runs that produced them.
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| weights | complete | 2026-08-21 07:43 | 48s | 1 |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · BF16 · 100,352×5120 |
| glitch surface | 110 undertrained, 85 plain-ASCII |
| lineage check | not checked (no claimed base model) |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/microsoft/phi-4)