microsoft/phi-1_5 warn
Glitch tokens that can silently corrupt ordinary input. Plus 1 minor note.
chat template: not found · view on Hugging Face ↗
Scan coverageStatic battery2026-08-25Weights battery2026-08-25Behavioral batterynot rundetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-25 |
| Weights battery | Weights forensics — no GPU, no download | complete 2026-08-2551,200-token embedding scanned · 1158 undertrained |
| Behavioral battery | Live-inference differentials | not run |
Ingot runs three batteries against a model. What each one checks →
Findings
Scanned 2026-08-25 · published from a community scan.
medium Undertrained (glitch) token surface in vocabulary
Embedding-norm scan flagged 1158 undertrained tokens (norm < 0.3× the vocabulary median of 1.245), including 157 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "TheNitrome", "guiIcon", "TheNitromeFan", "guiActive", "unfocusedRange", "channelAvailability", "GoldMagikarp", "srfN". 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.
low Undertrained tokens present; echo probe skipped (no chat template)
1158 undertrained tokens found (16 ASCII candidates) but the repo ships no chat template, so the behavioral echo probe was skipped. Treat the candidates as unverified risk surface.
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 2026-08-25.
| architecture | phi · 24 layers · 2048-dim |
| parameters | 1418.3M |
| vocabulary | 51,200 tokens |
| license | mit |
| serialization | safetensors |
| chat template | none |
| glitch-token surface | 1,158 undertrained candidates, 157 plain-ASCII |
Full measured fingerprint
| architectures | PhiForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 16 |
| revision | 77aa61eeac94 |
| HF snapshot | 56.8k downloads · 1.4k likes · updated 2025-11-24 · captured 2026-08-25 |
| embedding tensor | model.embed_tokens.weight · F16 · 51,200×2048 |
| embedding norms | median 1.2449 · mean 1.2054 |
| lineage check | no claimed base model |
| glitch-token samples | "TheNitrome", "guiIcon", "TheNitromeFan", "guiActive", "unfocusedRange", "channelAvailability", "GoldMagikarp", "srfN", "rawdownload", "Dragonbound", "Downloadha", "externalToEVAOnly" |
Battery runs (2)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| gpu | complete | 2026-08-25 21:47 | 33s | 1 |
| weights | complete | 2026-08-25 20:34 | 18s | 1 |
gpu run 2026-08-25 — measurements
| probes run | glitch |
| probes skipped | glitch-echo: no chat template |
weights run 2026-08-25 — measurements
| probes run | glitch-norm-scan, zero-template-token-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · F16 · 51,200×2048 |
| glitch surface | 1,158 undertrained, 157 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-1_5)