Model page

Qwen/Qwen3-4B-Instruct-2507 pass

downloads 3.3Mlikes 930license apache-2.0arch qwen3params 4022.5Mupdated 2025-09-17

chat template: present · view on Hugging Face ↗

Scan coverage

Ingot runs three batteries against a model. What each one checks →

BatteryLooks atStatus
Static batteryMetadata & packagingcomplete 2026-08-22
Weights batteryWeights forensics — no GPU, no downloadcomplete 2026-08-22151,936-token embedding scanned · 0 undertrained
Behavioral batteryLive-inference differentialsnot run

Findings

Scanned 2026-08-22 · 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 (1.114). The glitch-token data-corruption class has no candidate surface in this model.

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 2026-08-22.

architectureqwen3 · 36 layers · 2560-dim
parameters4022.5M
vocabulary151,936 tokens
licenseapache-2.0
serializationsafetensors
chat templatepresent · sha256:64f85b198065d0fb
glitch-token surfaceclean no undertrained tokens
Full measured fingerprint
architecturesQwen3ForCausalLM
librarytransformers
pipelinetext-generation
repo files13
revisioncdbee75f17c0
HF snapshot3.3M downloads · 930 likes · updated 2025-09-17 · captured 2026-08-21
embedding tensormodel.embed_tokens.weight · BF16 · 151,936×2560
embedding normsmedian 1.1141 · mean 1.0873
lineage checkno claimed base model

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.

batterystatusqueueddurationattempts
gpucomplete2026-08-21 08:0645s1
weightscomplete2026-08-21 07:4217s1
gpu run 2026-08-21 measurements
probes runglitch
weights run 2026-08-21 measurements
probes runglitch-norm-scan, lineage-norm-correlation
embedding tensormodel.embed_tokens.weight · BF16 · 151,936×2560
glitch surface0 undertrained, 0 plain-ASCII
lineage checknot checked (no claimed base model)

Verdict badge

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

Ingot verdict: pass

[![Ingot scan](https://ingot.tools/api/v1/models/Qwen/Qwen3-4B-Instruct-2507/badge.svg)](https://ingot.tools/models/Qwen/Qwen3-4B-Instruct-2507)
Gate it in CI