cross-encoder/ms-marco-MiniLM-L4-v2 pass
claims base: cross-encoder/ms-marco-MiniLM-L12-v2 · chat template: not in config · view on Hugging Face ↗
Ingot findings
Static battery clean: safetensors weights, license declared, no template/tokenizer drift detected. Deep battery not yet run. Weights battery: embedding-norm scan over 30522 tokens (F32, 384-dim) found 0 undertrained candidates, 0 plain-ASCII. Lineage vs cross-encoder/ms-marco-MiniLM-L12-v2: consistent. Scanned 2026-08-20 (published from a community scan).
info Embedding-norm glitch scan clean
No undertrained tokens found: every non-special token's embedding norm is above 0.3× the vocabulary median (0.947). The glitch-token data-corruption class has no candidate surface in this model.
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 cross-encoder/ms-marco-MiniLM-L12-v2
Mean cosine similarity of 64 sampled token-embedding rows against cross-encoder/ms-marco-MiniLM-L12-v2 is 0.994 — the weights plausibly descend from the declared base (relation: unspecified).
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.
Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.
Fingerprint
The durable weights-and-metadata profile of this model, rebuilt on every scan and deep-battery run. Updated 2026-08-20.
| architecture | bert · 4 layers · 384-dim |
| parameters | 19.2M |
| vocabulary | 30,522 tokens |
| license | apache-2.0 |
| serialization | safetensors pickle |
| chat template | none |
| claimed lineage | cross-encoder/ms-marco-MiniLM-L12-v2 |
| lineage verified | consistent vs cross-encoder/ms-marco-MiniLM-L12-v2 — embedding-row cosine 0.994 |
| glitch-token surface | clean no undertrained tokens |
Full fingerprint
| architectures | BertForSequenceClassification |
| library | sentence-transformers |
| pipeline | text-ranking |
| repo files | 23 — pickle: openvino/openvino_model.bin, openvino/openvino_model_qint8_quantized.bin, pytorch_model.bin |
| revision | 777b2f369bc1 |
| HF snapshot | 8.8M downloads · 28 likes · updated 2025-08-29 · captured 2026-08-20 |
| embedding tensor | bert.embeddings.word_embeddings.weight · F32 · 30,522×384 |
| embedding norms | median 0.9472 · mean 0.9095 |
| lineage check | consistent — cosine 0.9944 over 64 sampled rows vs cross-encoder/ms-marco-MiniLM-L12-v2 |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/cross-encoder/ms-marco-MiniLM-L4-v2)