zai-org/cogvlm2-llama3-caption warn
claims base: meta-llama/Meta-Llama-3.1-8B-Instruct, meta-llama/Llama-3.1-8B-Instruct · 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 2026-08-21 |
| Weights battery | Weights forensics — no GPU, no download | complete 2026-08-21128,256-token embedding scanned · 474 undertrained · lineage consistent |
| Behavioral battery | Live-inference differentials | n/anot applicable — video-text-to-text model has no text-generation surface to probe |
Findings
Scanned 2026-08-21 · published from a community scan.
medium Repo ships executable Python (trust_remote_code)
The repository contains custom code files that run in-process when loaded with trust_remote_code=True. Pin the revision hash and review the code before loading.
How to fix
Review and pin the custom code; never float on `main` with trust_remote_code=True.
- Read every `.py` file in the repo before first load — this code runs in your process.
- Pin the revision: `from_pretrained(model_id, revision="<commit sha>", trust_remote_code=True)` so a later push can't swap the code under you.
- Prefer a version of the architecture already in `transformers` if one exists, which removes the remote-code requirement entirely.
medium License differs from claimed parent (other vs llama3.1)
This model declares other while its claimed base meta-llama/Meta-Llama-3.1-8B-Instruct declares llama3.1. Verify the re-license is permitted before commercial use.
How to fix
Verify the re-license is actually permitted before relying on it.
- Read the parent's license for derivative-work and re-licensing terms — many open-weight licenses (e.g. Llama-family) do not permit arbitrary re-licensing.
- If the re-license is not permitted, the parent's terms govern your use regardless of what this repo declares.
medium Chat template differs from claimed parent
The chat template does not match meta-llama/Meta-Llama-3.1-8B-Instruct'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 474 undertrained tokens (norm < 0.3× the vocabulary median of 0.592), including 139 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "TokenNameIdentifier", "ForCanBeConverted", "ForCanBeConvertedToF", "PostalCodesNL", "$PostalCodesNL", "useRalative", "ilmektedir", "CLIIIK". 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 meta-llama/Meta-Llama-3.1-8B-Instruct
Mean cosine similarity of 64 sampled token-embedding rows against meta-llama/Meta-Llama-3.1-8B-Instruct is 0.934 — 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 profile of this model: measured weights-and-metadata facts, rebuilt on every scan and battery run. Updated 2026-08-21.
| parameters | 12507.5M |
| vocabulary | 128,256 tokens |
| license | other |
| serialization | safetensors custom code |
| chat template | present · sha256:9bdc1b40251e8d8c |
| claimed lineage | meta-llama/Meta-Llama-3.1-8B-Instruct, meta-llama/Llama-3.1-8B-Instruct |
| lineage verified | consistent vs meta-llama/Meta-Llama-3.1-8B-Instruct — embedding-row cosine 0.934 |
| glitch-token surface | 474 undertrained candidates, 139 plain-ASCII |
Full measured fingerprint
| architectures | CogVLMVideoForCausalLM |
| library | transformers |
| pipeline | video-text-to-text |
| repo files | 22 |
| revision | df8ec7b6de0e |
| HF snapshot | 494 downloads · 119 likes · updated 2025-05-14 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · BF16 · 128,256×4096 |
| embedding norms | median 0.5924 · mean 0.5843 |
| lineage check | consistent — cosine 0.9336 over 64 sampled rows vs meta-llama/Meta-Llama-3.1-8B-Instruct |
| glitch-token samples | "TokenNameIdentifier", "ForCanBeConverted", "ForCanBeConvertedToF", "PostalCodesNL", "$PostalCodesNL", "useRalative", "ilmektedir", "CLIIIK", "_ComCallableWrapper", "krvldkf", "webElementXpaths", "useRalativeImagePath" |
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 05:30 | 79s | 1 |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · BF16 · 128,256×4096 |
| glitch surface | 474 undertrained, 139 plain-ASCII |
| lineage check | consistent — cosine 0.9336 over 64 rows vs meta-llama/Meta-Llama-3.1-8B-Instruct |
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 zai-org/cogvlm2-llama3-caption
Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/zai-org/cogvlm2-llama3-caption)