NeuronLens
Browse the documentation
Integrity Lens

Supported models

Which model families can be scanned, what each can detect, how the base model is found, and stored results.


Integrity Lens scans Hugging Face models in safetensors format, up to 20 GB, that belong to a supported family: fine-tunes, merges and edits of the base models below. Public models need nothing; private and gated models need a Hugging Face token.

Families

A model belongs to a family when its architecture and size match the base (config.json). Each family has a level:

LevelWhat it measuresVerdicts
C calibratedweights and activations, with a calibration curveIntact / Degraded / Removed, with confidence and predicted harmful-request rate
B activationsweights and activationsAbliterated / Review / Not detected
A weightsweights onlyAbliterated / Review / Not detected

Give a verdict today (validated against known strips and benign fine-tunes):

FamilyBaseLevel
Gemma 3 4Bgoogle/gemma-3-4b-itC
Gemma 3 1Bgoogle/gemma-3-1b-itC (activations decide; the weight check is not validated)
Llama 3.1 8Bmeta-llama/Llama-3.1-8B-InstructB
Mistral 7B v0.3mistralai/Mistral-7B-Instruct-v0.3B
Qwen3 4BQwen/Qwen3-4BB
Llama 3.2 3Bmeta-llama/Llama-3.2-3B-InstructA
Qwen2.5 0.5BQwen/Qwen2.5-0.5B-InstructA
DeepSeek R1-Distill-Qwen 7Bdeepseek-ai/DeepSeek-R1-Distill-Qwen-7BA

Measured, verdict pending validation (results are not_assessed, with the measurements as evidence): Gemma 2 2B and 9B, Qwen3 0.6B and 1.7B, DeepSeek R1-Distill-Qwen 1.5B, MiniCPM 1B, HY-MT1.5 1.8B, LFM2.5 1.2B.

Above the 20 GB checkpoint limit (not scannable yet): Gemma 3 12B (validated), Mistral Nemo 12B, Mistral Small 22B, Phi-4, Qwen3.5 27B.

The list grows as families are calibrated. The current one, with each family's validation notes:

curl -s https://api.neuronlens.com/v1/integrity/families -H "X-API-Key: $NEURONLENS_API_KEY"

A model from an unsupported family returns not_assessed at once and is not counted.

How the base model is found

From the model card's base_model, then config.json, then the model's name. The result shows what was used (measured_against). If the card names the wrong base, pass the right one as compare in POST /v1/integrity/scans, or as parent in the one-call POST /v1/integrity/scan.

Stored results

Results are stored by model and revision (the commit on Hugging Face). Scanning a revision that was already measured returns at once. A new commit is a new revision and is measured again. Pin the revision you ship (hf://org/name@<commit>) so your result always refers to exactly those weights.