AI safety for engineers

A technical reading list for engineers who want to work on or near alignment: interpretability, adversarial robustness, red teaming, and the methods behind today's safety pipelines.

1. 1
   Training a Helpful and Harmless Assistant with RLHFAcademic Papers Advanced~2 hr read

   The engineering of an RLHF safety pipeline, end to end.

2. 2
   Red Teaming Language Models to Reduce HarmsAcademic Papers Advanced

   A repeatable methodology for finding model failures.

3. 3
   Discovering Latent Knowledge in Language Models Without SupervisionAcademic Papers Advanced

   An interpretability method aimed at detecting what a model 'believes'.

4. 4
   Robert Miles AI SafetyYouTube Beginner

   Concise technical explainers to fill conceptual gaps as you go.

5. 5
   LessWrongWebsites Intermediate

   Where much of the technical alignment discussion happens in the open.

See all learning paths →