Model page

deepseek-ai/DeepSeek-R1 warn

downloads 6.4Mlikes 13.6klicense mitarch deepseek_v3params 684531.4Mupdated 2025-03-27

chat template: present · view on Hugging Face ↗

Ingot findings

Static battery: 1 medium finding(s). Deep battery (behavioral differential, glitch-token pass) not yet run. Weights battery: embedding-norm scan over 129280 tokens (BF16, 7168-dim) found 607 undertrained candidates, 59 plain-ASCII. Scanned 2026-08-20 (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.

  1. Read every `.py` file in the repo before first load — this code runs in your process.
  2. 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.
  3. Prefer a version of the architecture already in `transformers` if one exists, which removes the remote-code requirement entirely.

medium Undertrained (glitch) token surface in vocabulary

Embedding-norm scan flagged 607 undertrained tokens (norm < 0.3× the vocabulary median of 3.196), including 59 plain-ASCII strings that can appear in ordinary input as identifiers — e.g. "kabungtor", "unisipyo", "ultatua", "pagklas", "jeftigelse", "bingkil", "nahimut", "asarangang". 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 GPU deep 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.

  1. 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.
  2. 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.
  3. 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.

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.

architecturedeepseek_v3 · 61 layers · 7168-dim
parameters684531.4M
vocabulary129,280 tokens
licensemit
serializationsafetensors custom code
chat templatepresent · sha256:563b1ce7d61d50a9
glitch-token surface607 undertrained candidates, 59 plain-ASCII
Full fingerprint
architecturesDeepseekV3ForCausalLM
librarytransformers
pipelinetext-generation
repo files174
revision56d4cbbb4d29
HF snapshot6.4M downloads · 13.6k likes · updated 2025-03-27 · captured 2026-08-20
embedding tensormodel.embed_tokens.weight · BF16 · 129,280×7168
embedding normsmedian 3.196 · mean 3.1572
lineage checkno claimed base model
glitch-token samples"kabungtor", "unisipyo", "ultatua", "pagklas", "jeftigelse", "bingkil", "nahimut", "asarangang", "Pagklas", "ugnawan", "Kadaghan", "unoang"

Verdict badge

Ship the verdict in your README — it always shows the latest published analysis:

Ingot verdict: warn

[![Ingot scan](https://ingot.tools/api/v1/models/deepseek-ai/DeepSeek-R1/badge.svg)](https://ingot.tools/models/deepseek-ai/DeepSeek-R1)
Gate it in CI