sentient-agi/OML-1.0-Fingerprinting

Health v2 · maintenance only

Adoption not part of the score

3497 stars · 232 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

A Python library for embedding secret cryptographic fingerprints into LLMs via fine-tuning, so model owners can prove ownership and detect unauthorized use. Fingerprints are special query-response pairs that uniquely identify a model and can be verified by owners or authorized users.

Use cases

• prove ownership of a fine-tuned llm
• detect unauthorized use of my model
• embed a secret signature into an llm
• verify which model i am talking to
• protect open-source models from theft
• fingerprint a model before releasing it

When to choose

• you distribute LLMs and need verifiable proof of ownership
• you want to detect if someone is using your model without permission
• you need to let authorized users verify they are talking to the genuine model

When to avoid

• you need general-purpose model watermarking of generated text rather than model fingerprinting
• you lack GPU resources for fine-tuning with DeepSpeed
• you just need standard model licensing or access control

Facets

library · maturity active

llm-training cryptography security machine-learning large-language-models artificial-intelligence security python model-fingerprinting model-ownership fine-tuning verifiable-ai model-provenance gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sentient-agi/OML-1.0-Fingerprinting")