Reconciling Symbols and Neural Networks

đŸ¤–đŸ§ NEW PAPERđŸ§ đŸ¤– What is the relationship between neural networks & symbolic systems? Based on evidence from mechanistic interpretability, we argue for a quasi-symbolic view of cognition: Neural nets implement symbols but only approximately Link: https://t.co/j8B3IOIMpe 1/n https://t.co/zpn6q00wqq

— Tom McCoy (@RTomMcCoy) October 14, 2026

For comparison, see Relational Nets Over Attractors, A Primer: Part 1, Design for a Mind, Version 3