jimouris/curl
Curl: Private LLMs through Wavelet-Encoded Look-Up Tables
This framework helps machine learning researchers evaluate large language models (LLMs) like GPT-2 or BERT while keeping the underlying data private and secure. It takes in trained LLMs and data, then processes them using secure multi-party computation to produce model evaluations without revealing sensitive information. This is ideal for ML researchers who need to work with confidential datasets.
No commits in the last 6 months.
Use this if you are a machine learning researcher who needs to evaluate large language models on sensitive data and require privacy-preserving techniques to protect that information.
Not ideal if you are looking for a production-ready system, as this is currently a research prototype.
Stars
16
Forks
4
Language
Python
License
MIT
Category
Last pushed
Apr 07, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/jimouris/curl"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
meta-pytorch/opacus
Training PyTorch models with differential privacy
tensorflow/privacy
Library for training machine learning models with privacy for training data
tf-encrypted/tf-encrypted
A Framework for Encrypted Machine Learning in TensorFlow
awslabs/fast-differential-privacy
Fast, memory-efficient, scalable optimization of deep learning with differential privacy
privacytrustlab/ml_privacy_meter
Privacy Meter: An open-source library to audit data privacy in statistical and machine learning...