Some thoughts on interpretability and a detailed description of my first real research project “TraDe”.

Why do I care so much about AI safety?