CMU Locus Lab

All

73 repositories

open-unlearning
Public
[NeurIPS D&B '25] The one-stop repository for LLM unlearning
open-source benchmarks right-to-be-forgotten
open-source benchmarks right-to-be-forgotten privacy-protection unlearning membership-inference-attacks membership-inference llms llm-privacy llm-unlearning
Python
•
MIT License
•141•502•9•3•Updated Mar 7, 2026Mar 7, 2026
llm-components
Public
Python
•0•3•0•0•Updated Feb 10, 2026Feb 10, 2026
mugrade
Public
grading scripts for mugrade.org
Python
•7•5•1•0•Updated Feb 7, 2026Feb 7, 2026
antidistillation-sampling
Public
Python
•
BSD 3-Clause "New" or "Revised" License
•4•15•2•0•Updated Nov 4, 2025Nov 4, 2025
safety-pretraining
Public
HTML
•0•3•0•0•Updated Sep 16, 2025Sep 16, 2025
confidential_llms
Public
Repository for the paper "Evaluating Language Model Reasoning about Confidential Information" and the PasswordEval benchmark.
Python
•0•2•0•0•Updated Aug 28, 2025Aug 28, 2025
llm-idiosyncrasies
Public
Code release for "Idiosyncrasies in Large Language Models"
Python
•
MIT License
•7•55•1•0•Updated Jul 21, 2025Jul 21, 2025
tofu
Public
Python
•
MIT License
•6•27•2•0•Updated Feb 25, 2025Feb 25, 2025
acr-memorization
Public
Python
•
BSD 3-Clause "New" or "Revised" License
•8•38•0•0•Updated Dec 19, 2024Dec 19, 2024
edge-of-stability
Public
Python
•24•74•1•0•Updated Dec 7, 2024Dec 7, 2024
context-parametric-inversion
Public
1•4•2•0•Updated Nov 17, 2024Nov 17, 2024
llava-token-compression
Public
Python
•3•47•0•0•Updated Nov 8, 2024Nov 8, 2024
ect
Public
Consistency Models Made Easy
Python
•12•327•9•1•Updated Oct 13, 2024Oct 13, 2024
diffusion-model-hallucination
Public
Python
•
MIT License
•6•54•1•0•Updated Sep 5, 2024Sep 5, 2024
qpth
Public
A fast and differentiable QP solver for PyTorch.
machine-learning deep-learning optimization
machine-learning deep-learning optimization pytorch quadratic-programming
Python
•
Apache License 2.0
•112•785•16•1•Updated Sep 3, 2024Sep 3, 2024
mpc.pytorch
Public
A fast and differentiable model predictive control (MPC) solver for PyTorch.
Python
•
MIT License
•163•1k•18•0•Updated Aug 27, 2024Aug 27, 2024
wanda
Public
A simple and effective LLM pruning approach.
network-pruning large-language-models
network-pruning large-language-models
Python
•
MIT License
•125•855•45•1•Updated Aug 9, 2024Aug 9, 2024
fast_adversarial
Public
[ICLR 2020] A repository for extremely fast adversarial training using FGSM
Python
•92•449•3•2•Updated Jul 25, 2024Jul 25, 2024
projected_sinkhorn
Public
Python
•15•88•1•1•Updated Jul 25, 2024Jul 25, 2024
robust_union
Public
[ICML'20] Multi Steepest Descent (MSD) for robustness against the union of multiple perturbation models.
Python
•4•25•0•1•Updated Jul 25, 2024Jul 25, 2024
lml
Public
The Limited Multi-Label Projection Layer
Python
•
MIT License
•6•59•0•1•Updated Jul 25, 2024Jul 25, 2024
scaling_laws_data_filtering
Public
Python
•4•64•0•0•Updated Apr 9, 2024Apr 9, 2024
massive-activations
Public
Code accompanying the paper "Massive Activations in Large Language Models"
large-language-models
large-language-models
Python
•
MIT License
•13•197•4•1•Updated Mar 4, 2024Mar 4, 2024
chatllm-vscode
Public
TypeScript
•
MIT License
•7•64•2•0•Updated Feb 4, 2024Feb 4, 2024
get
Public
Generative Equilibrium Transformer
Python
•
MIT License
•2•27•0•0•Updated Nov 11, 2023Nov 11, 2023
torchdeq
Public
Modern Fixed Point Systems using Pytorch
fixed-point deq equilibrium
fixed-point deq equilibrium fixed-point-method
Python
•
MIT License
•12•129•9•0•Updated Oct 31, 2023Oct 31, 2023
T-MARS
Public
Code for T-MARS data filtering
Python
•
MIT License
•5•35•1•0•Updated Aug 23, 2023Aug 23, 2023
FLYP
Public
Code for Finetune like you pretrain: Improved finetuning of zero-shot vision models
Python
•
MIT License
•14•106•6•0•Updated Aug 13, 2023Aug 13, 2023
smoothinv
Public
Single Image Backdoor Inversion via Robust Smoothed Classifiers
Python
•
MIT License
•1•17•1•0•Updated Jul 18, 2023Jul 18, 2023
tta_conjugate
Public
Test-Time Adaptation via Conjugate Pseudo-Labels
Python
•
MIT License
•3•42•1•0•Updated May 25, 2023May 25, 2023