Tag
GoodfireAI releases a research agenda on understanding neural geometry in language models, demonstrating the ability to precisely control a model's capabilities, such as removing its ability to speak German.
This article explores how sparse autoencoders (SAEs) can capture curved neural geometry, revealing three distinct ways SAE features represent manifolds, and presents an unsupervised pipeline to uncover geometric structure in neural representations.
Neural networks appear to speak English on the surface, but internally organize information in geometric space (curves, loops, surfaces, manifolds). Understanding "neural geometry" may be the key to understanding, debugging, and controlling models.
Goodfire AI announces a new research agenda focused on neural geometry to improve the understanding, debugging, and control of neural networks.
A research paper introducing Three-Phase Transformer (3PT), which applies Tesla's polyphase geometry to transformer architectures by organizing the residual stream into three 120° offset phases. The approach achieves 7.2% perplexity improvement on WikiText-103 with minimal parameters (0.00124% overhead) and 1.93× convergence speedup.