Beren's Projects
Quick experiments for testing effects of prompt length and generation size on GPT4 latency
A toolkit for developing and comparing reinforcement learning algorithms.
Hamiltonian MCMC implemented through predictive coding in a biologically plausible manner in the brain.
A number of tutorials I'm working through to learn haskell
This repository contains the code to implement the Hierarchical Actor-Critic (HAC) algorithm.
A super simple implementation of a hamiltonian MCMC scheme for tutorial purposes.
Colab notebook for my walkthrough on "Hopfield Networks are all you Need"
Github Pages template for academic personal websites, forked from mmistakes/minimal-mistakes
An image scraper implemented with Julia and Selenium
Code to reproduce experiments described in the Implementing Predictive Processing Report
A Python implementation of active inference for Markov Decision Processes
Repository for code for paper: "Backpropagation at the infinitesimal inference limit of energy based models: unifying predictive coding, equilibrium propagation, and contrastive hebbian learning"
Project for Hack the Burgh 19
Attempted replication of https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.231.5583&rep=rep1&type=pdf. Currently not working
some stuff
Testing a simple kalman filter implementation
An implementation (and partial replication) of the Lagrangian VAE paper.
Basic replication of the algorithms from https://wvvw.aaai.org/ojs/index.php/AAAI/article/view/4389
A collection of LaTeX templates and resources. Hopefully useful to someone out there (and myself)
Some work on whether human scanpaths are statistically Levy Flights
Some experiments and implementations of active inference for linear time invariant systems. Builds heavily on work by Manuel Baltieri and Sherin Grimbergen
A simple attempt to try building my own lisp dialect while also learning C
Animations and plots for LW post
Bot to automatically respond to each post on LW with a GPT4 comment
My work on autistic savantism and predictive processing from my master's dissertation
Code for the active inference simulations in: The Acquisition of Culturally Patterned Attention Styles under Active Inference
Implementation of Memorizing Transformers (ICLR 2022), attention net augmented with indexing and retrieval of memories using approximate nearest neighbors, in Pytorch
Model parallel transformers in JAX and Haiku