mechanistic interpretability

SALVE: New Technique Enables Mechanistic Control of Neural Networ

Researchers introduce SALVE, combining sparse autoencoders with latent vector editing for precise mechanistic control over neural network behaviors and outputs.

mechanistic interpretability

GPT-2 Dissected: How Transformer Layers Process Sentiment

New research reveals how GPT-2's layers divide labor between lexical and contextual processing during sentiment analysis, advancing our understanding of transformer internals.