Matthew A. Clarke's starred repositories
Awesome-Interpretability-in-Large-Language-Models
This repository collects all relevant resources about interpretability in LLMs
refusal_direction
Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".
MultiDimensionalFeatures
Code for reproducing our paper "Not All Language Model Features Are Linear"
neuroscope
Neuroscope:An Explainable AI Toolbox for Semantic Segmentation and Image Classification of Convolutional Neural Nets
PERCEPTION
Utilizing single-cell omics from patients tumor to predict response and resistance.
alphabetical_probe
Experimental code which trains 26 linear probes to detect the presence of alphabetic letters in GPT-J token strings, given their embeddings. Exploring the resulting vector arithmetic and its impact on GPT-J spelling abilities
pytest-action
:octocat: GitHub Action to run pytest with GitHub Job Summaries support :snake: :rocket:
Melanoma-LC
Network model .json file for the Langerhans Cell (LC), Melanoma and combined models first described in Howell, Davies et al., Science Advances 2023.
MutationOrder
MutationOrder is a tool to explore how the order in which mutations are acquired in an evolving cancer is constrained by the change in cell phenotypes that these mutations cause.
ASCETIC
ASCETIC (Agony-baSed Cancer EvoluTion InferenCe) is a novel framework for the inference of a set of statistically significant temporal patterns involving alternations in driver genes from cancer genomics data. Manuscript published at: https://www.nature.com/articles/s41467-023-41670-3
BI-BE-CS-183-2023
Introduction to Computational Biology and Bioinformatics Course at Caltech, 2023
nanoChatGPT
A crude RLHF layer on top of nanoGPT with Gumbel-Softmax trick
org-novelist
Org Novelist is a system for writing novel-length fiction using Emacs Org mode.
alpaca.cpp
Locally run an Instruction-Tuned Chat-Style LLM
SCEVAN
R package that automatically classifies the cells in the scRNA data by segregating non-malignant cells of tumor microenviroment from the malignant cells. It also infers the copy number profile of malignant cells, identifies subclonal structures and analyses the specific and shared alterations of each subpopulation.
ChatGPT3-Free-Prompt-List
A free guide for learning to create ChatGPT3 Prompts
tumor-prevention-rat-scRNAseq
Code and materials to support genomic analyses and reproduce results from the paper "Breast cancer prevention by short-term inhibition of TGFB signaling", published in Nature Communications in December 2022.