EOM0722/blogbook-gguf warn
Weights only ship in a format that can run code when loaded; EOS ids disjoint between config.json and generation_config.json; Chat template ends turns with <|eot_id|>, which is not a configured stop token. Plus 3 more issues.
claims base: beomi/Llama-3-Open-Ko-8B · chat template: present · view on Hugging Face ↗
Scan coverageStatic battery2026-08-27Weights battery2026-08-27Behavioral batterynot rundetails
| Battery | Looks at | Status |
|---|---|---|
| Static battery | Metadata & packaging | complete 2026-08-27 |
| Weights battery | Weights forensics: no GPU, no download | complete 2026-08-27128,256-token embedding scanned · 460 undertrained · lineage consistent · pickle audit clean |
| Behavioral battery | Live-inference differentials | not run |
Ingot runs three batteries against a model. What each one checks →
Findings
Scanned 2026-08-27 · published from a community scan.
medium Pickle-serialized weights, no safetensors
Weights ship only as pickle-based files (pytorch_model-00001-of-00017.bin, pytorch_model-00002-of-00017.bin, pytorch_model-00003-of-00017.bin, …). Loading pickle executes arbitrary code from the file — prefer a safetensors release or load in a sandbox.
How to fix
Convert the weights to safetensors before loading them anywhere that matters.
- Do not load the pickle files in-process — pickle deserialization executes arbitrary code from the file.
- Convert locally in a sandbox: `pip install safetensors` and use `safetensors.torch.save_file` on a state dict loaded with `torch.load(..., weights_only=True)` (refuses most code-execution payloads), or use Hugging Face's `convert.py` space/script.
- Pin the exact revision hash you converted from, and load only your converted safetensors artifact from then on.
medium EOS ids disjoint between config.json and generation_config.json
config.json declares eos_token_id [128001] while generation_config.json declares [2] with no overlap. Runtimes read one or the other, so at least one of them stops generation on the wrong token (or never). Align both files on the token the chat template actually ends turns with.
How to fixingot patch
Align the stop-token declarations — a pure metadata fix to `generation_config.json` (and `config.json`).
- Identify the token the chat template actually ends assistant turns with (e.g. `<|eot_id|>`, `<end_of_turn>`, `<|im_end|>`) and make sure its id is in `generation_config.json`'s `eos_token_id` list.
- Keep `config.json`'s `eos_token_id` consistent with (or a subset of) `generation_config.json`'s — runtimes differ in which file they read.
- For the pad-equals-EOS hazard: repoint `pad_token` at a dedicated padding token before fine-tuning; serving is unaffected.
- Until the repo is fixed, pass explicit stop tokens to your serving stack (e.g. vLLM `stop_token_ids`, llama.cpp `--override-kv tokenizer.ggml.eos_token_id`).
medium Chat template ends turns with <|eot_id|>, which is not a configured stop token
The chat template terminates assistant turns with <|eot_id|>, but the effective EOS set (config.json ∪ generation_config.json = [128001,2] → ["<|end_of_text|>"]) never stops on it. Config-honoring runtimes generate past the terminator until the token budget is exhausted — runaway cost and self-continuing fake turns. Add <|eot_id|>'s id to generation_config.json's eos_token_id.
How to fixingot patch
Align the stop-token declarations — a pure metadata fix to `generation_config.json` (and `config.json`).
- Identify the token the chat template actually ends assistant turns with (e.g. `<|eot_id|>`, `<end_of_turn>`, `<|im_end|>`) and make sure its id is in `generation_config.json`'s `eos_token_id` list.
- Keep `config.json`'s `eos_token_id` consistent with (or a subset of) `generation_config.json`'s — runtimes differ in which file they read.
- For the pad-equals-EOS hazard: repoint `pad_token` at a dedicated padding token before fine-tuning; serving is unaffected.
- Until the repo is fixed, pass explicit stop tokens to your serving stack (e.g. vLLM `stop_token_ids`, llama.cpp `--override-kv tokenizer.ggml.eos_token_id`).
medium License differs from claimed parent (apache-2.0 vs other)
This model declares apache-2.0 while its claimed base beomi/Llama-3-Open-Ko-8B declares other. 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 0.601), 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.
medium Chat template uses tokens with all-zero embeddings
3 special token(s) the chat template emits have effectively-zero embedding rows (norm < 0.000001) — never trained: <|start_header_id|> (128006), <|end_header_id|> (128007), <|eot_id|> (128009). Serving with this template feeds the model no-op inputs at turn boundaries, and fine-tuning on it produces NaN gradients or silently broken checkpoints (the Llama-3 base-model incident). Initialize these rows (e.g. to the embedding mean) before fine-tuning, or use a base-model prompt format instead of the chat template.
info Pickle static analysis clean
Opcode-level parse of pytorch_model-00001-of-00017.bin, pytorch_model-00002-of-00017.bin, pytorch_model-00003-of-00017.bin, pytorch_model-00004-of-00017.bin (no code executed) found only standard serialization globals (3 distinct, all torch/collections/numpy). Pickle remains an executable format — this verifies the current bytes, not future uploads; prefer a safetensors release.
info Weights consistent with claimed parent beomi/Llama-3-Open-Ko-8B
Mean cosine similarity of 64 sampled token-embedding rows against beomi/Llama-3-Open-Ko-8B is 1.000 — the weights plausibly descend from the declared base (relation: unspecified).
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 EOM0722/blogbook-gguf
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 2026-08-27.
| architecture | llama · 32 layers · 4096-dim |
| vocabulary | 128,256 tokens |
| license | apache-2.0 |
| serialization | no safetensors pickle |
| chat template | present · sha256:ba03a121d097859c |
| claimed lineage | beomi/Llama-3-Open-Ko-8B |
| lineage verified | consistent vs beomi/Llama-3-Open-Ko-8B — embedding-row cosine 1.000 |
| glitch-token surface | 460 undertrained candidates, 131 plain-ASCII |
Full measured fingerprint
| architectures | LlamaForCausalLM |
| library | transformers |
| pipeline | text-generation |
| repo files | 25 — pickle: pytorch_model-00001-of-00017.bin, pytorch_model-00002-of-00017.bin, pytorch_model-00003-of-00017.bin, pytorch_model-00004-of-00017.bin, pytorch_model-00005-of-00017.bin, pytorch_model-00006-of-00017.bin, pytorch_model-00007-of-00017.bin, pytorch_model-00008-of-00017.bin, pytorch_model-00009-of-00017.bin, pytorch_model-00010-of-00017.bin |
| revision | bf60869f3dab |
| HF snapshot | 13 downloads · 0 likes · updated 2024-08-04 · captured 2026-08-25 |
| pickle audit | pytorch_model-00001-of-00017.bin, pytorch_model-00002-of-00017.bin, pytorch_model-00003-of-00017.bin, pytorch_model-00004-of-00017.bin — 3 standard global(s) |
| embedding tensor | model.embed_tokens.weight · F16 · 128,256×4096 |
| embedding norms | median 0.6007 · mean 0.5907 |
| lineage check | consistent — cosine 1 over 64 sampled rows vs beomi/Llama-3-Open-Ko-8B |
| glitch-token samples | "TokenNameIdentifier", "ForCanBeConverted", "ForCanBeConvertedToF", "PostalCodesNL", "$PostalCodesNL", "useRalative", "ilmektedir", "CLIIIK", "_ComCallableWrapper", "krvldkf", "webElementXpaths", "useRalativeImagePath" |
Battery runs (1)the run trace behind the findings — what each job measured
| battery | status | queued | duration | attempts |
|---|---|---|---|---|
| weights | complete | 2026-08-25 22:35 | 2m | 1 |
weights run 2026-08-25 — measurements
| probes run | glitch-norm-scan, zero-template-token-scan, pickle-static-analysis, lineage-norm-correlation |
| embedding tensor | model.embed_tokens.weight · F16 · 128,256×4096 |
| glitch surface | 460 undertrained, 131 plain-ASCII |
| lineage check | consistent — cosine 1 over 64 rows vs beomi/Llama-3-Open-Ko-8B |
Verdict badge
Ship the verdict in your README — it always shows the latest published analysis:
[](https://ingot.tools/models/EOM0722/blogbook-gguf)