Qwen/Qwen2.5-14B-Instruct warn
The chat template differs from its base model, which changes behavior; glitch tokens that can silently corrupt ordinary input.
claims base: Qwen/Qwen2.5-14B · chat template: present · view on Hugging Face ↗
Scan coverageStatic battery2026-08-22Weights battery2026-08-22Behavioral batteryfaileddetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-22 |
| Weights battery | Weights forensics: no GPU, no download | complete 2026-08-22152,064-token embedding scanned · 5737 undertrained · lineage consistent |
| Behavioral battery | Live-inference differentials | failedTraceback (most recent call last): | raise HTTPStatusError(message, request=request, response=self) | httpx.HTTPStatusError: Client error '429 Too Many Requests' for url 'https://huggingface.co/api/models/Qwen/Qwen2.5-14B-Instruct' | Traceback (most recent call last): | raise _format(HfHubHTTPError, message, response) from e | huggingface_hub.errors.HfHubHTTPError: (Request ID: Root=1-6a8e1647-1f0689945d909ed00b4f7bed;ad7773dc-ef8c-464f-b4df-6b2f206e2d99) |
Ingot runs three batteries against a model. What each one checks →
Findings
Scanned 2026-08-22 · published from a community scan.
medium Chat template differs from claimed parent
The chat template does not match Qwen/Qwen2.5-14B's. Template drift silently changes model behavior even when weights are identical — 37% of drifted derivatives in our census left it undisclosed. Diff the templates before deploying.
How to fixingot patch
Restore the parent's chat template in `tokenizer_config.json` — a pure metadata fix.
- Run `ingot patch <owner/model>` — the patch manifest carries the parent's template and applies it to a local copy's `tokenizer_config.json`.
- Or fix by hand: copy the `chat_template` value from the parent repo's `tokenizer_config.json` into this model's, and pin your serving stack to that file.
- If the drift was intentional (the author retrained on a new template), confirm that in the model card before "fixing" it — restoring the parent template on retrained weights changes behavior too.
medium Undertrained (glitch) token surface in vocabulary
Embedding-norm scan flagged 5737 undertrained tokens (norm < 0.3× the vocabulary median of 1.225), including 119 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "<unk>", "ForCanBeConverted", "ForCanBeConvertedToF", "$PostalCodesNL", "PostalCodesNL", "(stypy", "TokenNameIdentifier", "useRalative". 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.
info Weights consistent with claimed parent Qwen/Qwen2.5-14B
Mean cosine similarity of 64 sampled token-embedding rows against Qwen/Qwen2.5-14B is 1.000 — the weights plausibly descend from the declared base (relation: unspecified).
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.
Fix it
Some findings are metadata-level and patchable — apply the fixes to your local copy (your weights never leave your machine):
npx @ingotai/scan patch Qwen/Qwen2.5-14B-Instruct
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 | qwen2 · 48 layers · 5120-dim |
| parameters | 14770.0M |
| vocabulary | 152,064 tokens |
| license | apache-2.0 |
| serialization | safetensors |
| chat template | present · sha256:cd8e9439f0570856 |
| claimed lineage | Qwen/Qwen2.5-14B |
| lineage verified | consistent vs Qwen/Qwen2.5-14B — embedding-row cosine 1.000 |
| glitch-token surface | 5,737 undertrained candidates, 119 plain-ASCII |
Full measured fingerprint
| architectures | Qwen2ForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 18 |
| revision | cf98f3b3bbb4 |
| HF snapshot | 2.8M downloads · 359 likes · updated 2024-09-25 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · BF16 · 152,064×5120 |
| embedding norms | median 1.2252 · mean 1.1403 |
| lineage check | consistent — cosine 0.9999 over 64 sampled rows vs Qwen/Qwen2.5-14B |
| glitch-token samples | "<unk>", "ForCanBeConverted", "ForCanBeConvertedToF", "$PostalCodesNL", "PostalCodesNL", "(stypy", "TokenNameIdentifier", "useRalative", "useRal", "thuisontvangst", "Cumhurba", "NdrFc" |
Battery runs (2)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| gpu | complete | 2026-08-21 21:32 | 2m | 1 |
| weights | complete | 2026-08-21 07:42 | 66s | 1 |
gpu run 2026-08-21 — measurements
| probes run | glitch |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · BF16 · 152,064×5120 |
| glitch surface | 5,737 undertrained, 119 plain-ASCII |
| lineage check | consistent — cosine 0.9999 over 64 rows vs Qwen/Qwen2.5-14B |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/Qwen/Qwen2.5-14B-Instruct)