Authors
Loading...
Survey reviews mechanistic interpretability approaches in neural networks, suggesting a path for deeper AI understanding.
Kowalska et al. (2026) studied this question.