Fast and Easy Infinite Neural Networks in Python

CVPR 2024-Improved Implicit Neural Representation with Fourier Reparameterized Training

ICML2025-Inductive Gradient Adjustment for Spectral Bias in Implicit Neural Representations

Existing literature about training-data analysis.

A unified framework for attributing model components, data, and training dynamics to model behavior.

Official repository for "FOCUS: First Order Concentrated Updating Scheme"

Code for "What Happens During the Loss Plateau? Understanding Abrupt Learning in Transformers" (NeurIPS 2025)

Code for 'Towards a Theoretical Understanding of the 'Reversal Curse' via Training Dynamics'

Source code for <Probability Consistency in Large Language Models: Theoretical Foundations Meet Empirical Discrepancies>

Code for "Effect of equivariance on training dynamics"

Official repository for the EMNLP 2024 paper "How Hard is this Test Set? NLI Characterization by Exploiting Training Dynamics"

Code for "Abrupt Learning in Transformers: A Case Study on Matrix Completion" (NeurIPS 2024)

ASPIRE Scalar Vortex Visualizer — .NET MAUI desktop app for comparing ML inference runs with scientific rigor.

Code for tracking concept emergence via attention-head binding (EB*). Pythia experiments across 160M-2.8B parameters.

Track how discrete representations evolve during neural network training — lifecycle events, phase transitions, ontology discovery, and causal verification

Atomic benchmark suite showing drift can act as an early warning before direct symmetry detection in gradual-breaking regimes, with reversal controls, finite-budget sensitivity tests, and exact alarm-time validation.

Reimplementation of the Sliced Information Plane (SIP) framework from Wongso, Ghosh, and Motani (2025) for analyzing deep neural network training dynamics. The repo uses Sliced Mutual Information (SMI) to obtain scalable, finite dependence estimates in high‑dimensional, deterministic settings, and applies them to MNIST MLP experiments.

Supplementary code for the paper 'Dynamic Rescaling for Training GNNs' to be published at NeurIPS 2024

A research project investigating how LSTM training dynamics relate to dynamical stability and order–chaos transitions through Finite-Time Lyapunov Exponent (FTLE) analysis.

This repository documents the process of finding the optimal learning rate for deep neural networks