Please note: This curriculum is (almost) a copy of https://bluedot.org/courses/alignment’*s Alignment curriculum, for ease of access. This fellowship is not a Bluedot initiative.*

TAISI Fellowship Readings

Readings

Week 1 (Intro to LLMs)

Week 2 (Intro to AI Alignment)

Week 3 (Reinforcement Learning from Human (or AI) Feedback)

Week 4 (Scalable Oversight)

Week 5 (Robustness unlearning and control)

Week 6 (Mechanistic Interpretability)

Week 7 (Technical governance approaches)

Week 8 (Contributing to AI Safety 🥳)