Abulhair Saparov

This article analyzes how language models understand and infer causal relationships.

An overview of mechanistic interpretability in transformer-based language models.

Researchers investigate how transformers can improve their search abilities using training techniques.