RefalMachine/RuadaptQwen3-4B-Hybrid warn
The chat template differs from its base model, which changes behavior; its tokenizer differs from its claimed base model. Plus 1 minor note.
claims base: Qwen/Qwen3-4B · chat template: present · view on Hugging Face ↗
Scan coverageStatic battery2026-08-22Weights battery2026-08-22Behavioral batterynot rundetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-22 |
| Weights battery | Weights forensics: no GPU, no download | complete 2026-08-22146,260-token embedding scanned · 0 undertrained · lineage inconclusive |
| Behavioral battery | Live-inference differentials | not run |
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/Qwen3-4B'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 Vocabulary size differs from claimed parent (146260 vs 151936)
A changed vocab means changed tokenization: strings will split differently than on Qwen/Qwen3-4B, which can shift behavior on identifiers, codes, and non-English text.
How to fixweight-level
Not patchable: the vocab size mirrors the embedding matrix in the weights. Verify the change was intentional.
- Do not edit `vocab_size` in config.json to "match the parent" — it must equal the embedding table in the shipped weights or the model won't load.
- Diff the tokenizers (`tokenizer.json` / added_tokens) against the parent to see what was added or removed, and test your own identifiers, codes, and non-English text through both.
- If the drift is unexplained by the model card, treat tokenization-sensitive behavior as unvalidated on this model.
info Embedding-norm glitch scan clean
No undertrained tokens found: every non-special token's embedding norm is above 0.3× the vocabulary median (1.142). The glitch-token data-corruption class has no candidate surface in this model.
low Lineage vs claimed parent Qwen/Qwen3-4B inconclusive
Mean embedding-row cosine similarity to the declared base is 0.697 — below the 0.8 typical of true derivatives but not low enough to call mislabeled. Heavy continued pretraining or vocabulary surgery can look like this; verify provenance before relying on the parent's safety or licensing posture.
How to fix
Fix or verify the `base_model` declaration so lineage checks can run.
- If you own the repo: correct the `base_model` field in the model card metadata to the real, public parent.
- If you don't: identify the true parent (config architecture + weight shapes narrow it fast) and re-scan with that lineage in mind.
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 RefalMachine/RuadaptQwen3-4B-Hybrid
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 | qwen3 · 36 layers · 2560-dim |
| parameters | 4007.9M |
| vocabulary | 146,260 tokens |
| license | apache-2.0 |
| serialization | safetensors |
| chat template | present · sha256:87a2728cb8dc9fe4 |
| claimed lineage | Qwen/Qwen3-4B |
| lineage verified | inconclusive vs Qwen/Qwen3-4B — embedding-row cosine 0.697 |
| glitch-token surface | clean no undertrained tokens |
Full measured fingerprint
| architectures | Qwen3ForCausalLM |
| repo files | 13 |
| revision | bb890732923b |
| HF snapshot | 352 downloads · 4 likes · updated 2025-08-26 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · BF16 · 146,260×2560 |
| embedding norms | median 1.1422 · mean 1.1266 |
| lineage check | inconclusive — cosine 0.6972 over 64 sampled rows vs Qwen/Qwen3-4B |
Battery runs (1)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| weights | complete | 2026-08-21 05:32 | 58s | 1 |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · BF16 · 146,260×2560 |
| glitch surface | 0 undertrained, 0 plain-ASCII |
| lineage check | inconclusive — cosine 0.6972 over 64 rows vs Qwen/Qwen3-4B |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/RefalMachine/RuadaptQwen3-4B-Hybrid)