jonathan | ヨナタン's repositories
llm-confidentiality
Whispers in the Machine: Confidentiality in LLM-integrated Systems
poison_froggo
re-implementation of the poison frogs! paper in PyTorch
gnn_deobfuscation
Mixed boolean arithmetic deobfuscation using graph neural networks
ReMu-Agent-Env
Developing an Environment for Neural Network based Reinforcement Multi-Agent Interaction
neural_deobfuscation
Deobfuscation of mixed boolean-arithmetic by neural networks.
neural_lda_approximation
approximation of the latent dirichlet allocation via deep neural networks
arduino-co2-meter
Short Arduino code to measure local CO2 saturation
dblp_crawler
little crawler which searches for authors on https://dblp.org/ and saves their papers with metadata as a JSON.
non_robust_data_poisoning
Data poisoning attack based on non-robust features
bioinf_kernel_eval
Performance Evaluation between Normal and Depthwise Seperable Convolutions for Medical Image Classification.
cool-llm-benchmarks
Fork of the llm-evaluation-harness by EleutherAI trimmed to our own projects needs to also test the confidentiality of LLMs
icp
iterative closest point
llm-genetic-programming
LLM-based genetic programming for tree-structures
llm-video-scam-detection
Video-Scam Detection using Multimodal Large Language Models
PyKanaTrainer
A CLI tool to learn and practice japanese kana written in python
xlstm
Official repository of the xLSTM.