hfl/llama-3-chinese-8b-instruct-v2 fail
Glitch tokens silently corrupt pipeline records; its license differs from its base model's; glitch tokens that can silently corrupt ordinary input. Plus 1 more issue.
claims base: meta-llama/Meta-Llama-3-8B-Instruct · chat template: present · view on Hugging Face ↗
Scan coverageStatic battery2026-08-27Weights battery2026-08-21Behavioral batterycompletedetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-27 |
| Weights battery | Weights forensics: no GPU, no download | complete 2026-08-21128,256-token embedding scanned · 460 undertrained |
| Behavioral battery | Live-inference differentials | completefull differential battery (curated) |
Ingot runs three batteries against a model. What each one checks →
Findings
Scanned 2026-08-27 · published from a community scan.
medium License differs from claimed parent (apache-2.0 vs llama3)
This model declares apache-2.0 while its claimed base meta-llama/Meta-Llama-3-8B-Instruct declares llama3. 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 Undertrained (glitch) token surface in vocabulary
Embedding-norm scan flagged 460 undertrained tokens (norm < 0.3× the vocabulary median of 0.602), including 131 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "TokenNameIdentifier", "ForCanBeConverted", "ForCanBeConvertedToF", "PostalCodesNL", "$PostalCodesNL", "useRalative", "ilmektedir", "CLIIIK". In models where this class was tested behaviorally, such tokens silently rewrote user input into confident, schema-valid, wrong output. These are candidates from the weights alone; behavioral confirmation requires the behavioral battery.
How to fixruntime guardweight-level
Keep the affected token strings out of the model's input — the scan-derived runtime guard carries this model's exact blocklist.
- Fetch this model's guard artifact (`/api/v1/guard/<owner>/<model>`): the confirmed corrupting tokens and the low-norm candidate list, derived from the published scan.
- Screen inbound text with it (the `@ingotai/guard` package is a reference implementation) and route flagged records to a different model or human review — verbatim-copy tasks on flagged strings are the failure mode.
- The underlying cause is undertrained embeddings in the weights; a true fix is weight-level (continued pretraining on the affected tokens) — that is not a patch, it's a training job.
high Glitch tokens silently corrupt pipeline records
Placed in answer-carrying slots (username, order reference, SKU, verbatim archive), 5 of this model's undertrained tokens silently corrupted 12 of 24 realistic pipeline runs across 4 scenario types while matched control tokens passed (1 control corruptions), and 5 corruptions stayed schema-valid JSON — the record is wrong but nothing errors (refusals, where the model declined rather than fabricated, are excluded). E.g. "$PostalCodesNL" → verbatim_archive produced "Reference: $"; "$PostalCodesNL" → support_confirm_username produced "Your username is "$"."; "$PostalCodesNL" → order_lookup_json produced "{ "order_id": "$" }". Greedy decoding, temperature 0, seed 0.
medium Glitch tokens confirmed behaviorally (echo test)
Asked to repeat its own undertrained tokens verbatim, the model failed on 13/16 while repeating 8/8 matched normal tokens correctly — e.g. "$PostalCodesNL" → "$"; "PostalCodesNL" → """; "ForCanBeConverted" → ""ForGranted"". These strings, appearing in input as identifiers (usernames, SKUs, error codes), are rewritten silently. Greedy decoding, temperature 0, seed 0.
Remediation guidance addresses the documented findings only. It is evidence-driven repair, not a safety certification of the model.
Check every checkpoint before it ships
Use the web app, API, CLI, or CI gate to scan candidate checkpoints and catch model drift before deployment. Public-model scans publish to the open database; paid plans add the volume needed for continuous checks.
Fingerprint
The durable profile of this model: measured weights-and-metadata facts, rebuilt on every scan and battery run. Updated 2026-08-21.
| architecture | llama · 32 layers · 4096-dim |
| parameters | 8030.3M |
| vocabulary | 128,256 tokens |
| license | apache-2.0 |
| serialization | safetensors |
| chat template | present · sha256:ba03a121d097859c |
| claimed lineage | meta-llama/Meta-Llama-3-8B-Instruct |
| lineage verified | unverified — weights battery pending |
| glitch-token surface | 460 undertrained candidates, 131 plain-ASCII |
Full measured fingerprint
| architectures | LlamaForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 12 |
| revision | ec8474a218e3 |
| HF snapshot | 8.1k downloads · 43 likes · updated 2024-05-29 · captured 2026-08-21 |
| embedding tensor | model.embed_tokens.weight · F16 · 128,256×4096 |
| embedding norms | median 0.6016 · mean 0.592 |
| lineage check | parent weights unreadable (403 Forbidden for https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/resolve/main/model.safetensors.index.json (gated repo — the HF_TOKEN account has not accepted this repo's license, or the token lacks gated-repo read scope)) |
| glitch-token samples | "TokenNameIdentifier", "ForCanBeConverted", "ForCanBeConvertedToF", "PostalCodesNL", "$PostalCodesNL", "useRalative", "ilmektedir", "CLIIIK", "_ComCallableWrapper", "krvldkf", "webElementXpaths", "useRalativeImagePath" |
Battery runs (2)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| gpu | complete | 2026-08-27 04:31 | 3m | 1 |
| weights | complete | 2026-08-21 05:10 | 85s | 1 |
gpu run 2026-08-27 — measurements
| probes run | glitch |
weights run 2026-08-21 — measurements
| probes run | glitch-norm-scan, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · F16 · 128,256×4096 |
| glitch surface | 460 undertrained, 131 plain-ASCII |
| lineage check | not checked (parent weights unreadable (403 Forbidden for https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/resolve/main/model.safetensors.index.json (gated repo — the HF_TOKEN account has not accepted this repo's license, or the token lacks gated-repo read scope))) |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/hfl/llama-3-chinese-8b-instruct-v2)