Skip to content

Advanced

Why models get things wrong

Hallucinations, sycophancy, bias, misleading explanations: where model errors really come from, what research has seen by looking inside, and how to protect yourself when using or evaluating them.

4 lessons~39 mina quiz with every lesson

Start the course in the Academy

The lessons of the course

  1. Hallucination, an error that training encourages10 min
  2. Sycophancy: when it tells you you're right9 min
  3. Les biais : un reflet déformé des données10 min
  4. Regarder à l'intérieur : ce que l'interprétabilité a montré10 min

Every lesson can be read here in full; the quiz, your progress and the certificate are in the Academy.

Other advanced courses

The whole Academy →The AI glossary