zai-org/GLM-4.7-Flash 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 2026-08-22154,880-token embedding scanned · 1003 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 1003 undertrained tokens (norm < 0.3× the vocabulary median of 0.774), including 122 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "$PostalCodesNL", "_typeDefinitionSize", "_ComCallableWrapper", "ForCanBeConvertedToF", "PostalCodesNL", "ForCanBeConverted", "estattet", "irsiniz". 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 | glm4_moe_lite · 47 layers · 2048-dim |
| parameters | 31221.5M |
| vocabulary | 154,880 tokens |
| license | mit |
| serialization | safetensors |
| chat template | none |
| glitch-token surface | 1,003 undertrained candidates, 122 plain-ASCII |
Full measured fingerprint
| architectures | Glm4MoeLiteForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 58 |
| revision | 7dd20894a642 |
| HF snapshot | 2.0M downloads · 1.8k likes · updated 2026-01-29 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · BF16 · 154,880×2048 |
| embedding norms | median 0.7745 · mean 0.7555 |
| lineage check | no claimed base model |
| glitch-token samples | "$PostalCodesNL", "_typeDefinitionSize", "_ComCallableWrapper", "ForCanBeConvertedToF", "PostalCodesNL", "ForCanBeConverted", "estattet", "irsiniz", "ociazione", "unehmen", "_FieldOffsetTable", "_AdjustorThunk" |
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 07:42 | 23s | 1 |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · BF16 · 154,880×2048 |
| glitch surface | 1,003 undertrained, 122 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/zai-org/GLM-4.7-Flash)