Model page

Qwen/Qwen3-14B warn

The chat template differs from its base model, which changes behavior; glitch tokens that can silently corrupt ordinary input.

downloads 3.9Mlikes 489license apache-2.0arch qwen3params 14768.3Mupdated 2025-07-26

claims base: Qwen/Qwen3-14B-Base · chat template: present · view on Hugging Face ↗

Scan coverageStatic battery2026-08-22Weights battery2026-08-22Behavioral batteryfaileddetails
BatteryLooks atStatus
Static batteryMetadata & packagingcomplete 2026-08-22
Weights batteryWeights forensics: no GPU, no downloadcomplete 2026-08-22151,936-token embedding scanned · 3139 undertrained · lineage consistent
Behavioral batteryLive-inference differentialsfailed raise _format(HfHubHTTPError, message, response) from e | huggingface_hub.errors.HfHubHTTPError: (Request ID: Root=1-6a8e164c-49a503ff20aa4ce34a3310f5;33f993f6-c251-4cb9-9761-9e654b464e64) | Traceback (most recent call last): | raise LocalEntryNotFoundError( | huggingface_hub.errors.LocalEntryNotFoundError: Got: HfHubHTTPError: (Request ID: Root=1-6a8e164c-49a503ff20aa4ce34a3310f5;33f993f6-c251-4cb9-9761-9e654b464e64) | An error happened while trying to locate the files on the Hub, and w

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/Qwen3-14B-Base'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.

  1. Run `ingot patch <owner/model>` — the patch manifest carries the parent's template and applies it to a local copy's `tokenizer_config.json`.
  2. 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.
  3. 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 3139 undertrained tokens (norm < 0.3× the vocabulary median of 1.371), including 107 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "$PostalCodesNL", "PostalCodesNL", "ForCanBeConvertedToF", "ForCanBeConverted", "useRalative", "useRal", "thuisontvangst", "sexkontakte". 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.

  1. 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.
  2. 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.
  3. 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/Qwen3-14B-Base

Mean cosine similarity of 64 sampled token-embedding rows against Qwen/Qwen3-14B-Base is 0.998 — the weights plausibly descend from the declared base (relation: unspecified).

Put this result in your workflow

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/Qwen3-14B

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.

architectureqwen3 · 40 layers · 5120-dim
parameters14768.3M
vocabulary151,936 tokens
licenseapache-2.0
serializationsafetensors
chat templatepresent · sha256:a55ee1b1660128b7
claimed lineageQwen/Qwen3-14B-Base
lineage verifiedconsistent vs Qwen/Qwen3-14B-Base — embedding-row cosine 0.998
glitch-token surface3,139 undertrained candidates, 107 plain-ASCII
Full measured fingerprint
architecturesQwen3ForCausalLM
librarytransformers
pipelinetext-generation
repo files18
revision40c069824f42
HF snapshot2.1M downloads · 444 likes · updated 2025-07-26 · captured 2026-08-21
embedding tensormodel.embed_tokens.weight · BF16 · 151,936×5120
embedding normsmedian 1.3711 · mean 1.3067
lineage checkconsistent — cosine 0.9978 over 64 sampled rows vs Qwen/Qwen3-14B-Base
glitch-token samples"$PostalCodesNL", "PostalCodesNL", "ForCanBeConvertedToF", "ForCanBeConverted", "useRalative", "useRal", "thuisontvangst", "sexkontakte", "webElementX", "NdrFc", "sextreffen", "swingerclub"
Battery runs (1)the run trace behind the findings — what each job measured
batterystatusqueueddurationattempts
weightscomplete2026-08-21 07:4262s1
weights run 2026-08-21 — measurements
probes runglitch-norm-scan, lineage-norm-correlation
embedding tensormodel.embed_tokens.weight · BF16 · 151,936×5120
glitch surface3,139 undertrained, 107 plain-ASCII
lineage checkconsistent — cosine 0.9978 over 64 rows vs Qwen/Qwen3-14B-Base

Verdict badge

Ship the verdict in your README — it always shows the latest published analysis:

Ingot verdict: warn

[![Ingot scan](https://ingot.tools/api/v1/models/Qwen/Qwen3-14B/badge.svg)](https://ingot.tools/models/Qwen/Qwen3-14B)
Gate it in CI