hAIgopala › AI terms › Rapid Fire (3 terms per short) › Day 16

Constitutional AI / RLAIF

Generative AI & LLMsChapter 4: Choosing the LLMTerm 77 of 253

Aligning a model using a set of written principles and AI feedback instead of only human ratings.

At Sunrise School

Gopal asks

Can a model be aligned with written principles?

Bit answers

Yes. Constitutional AI uses a written set of principles, like “be helpful, be harmless”, and AI feedback to critique and improve answers.

In the real world

Model critiques itself against a rulebook.

Watch the short

Short coming soonSeason 5 · Day 16Follow @haigopala

Taught together with

Instruction Tuning · Scaling Laws

Practise it

Try this term and all 253 others in the free Model Training Arena.

Open the Arena →

Follow Bit daily: YouTube English · తెలుగు · हिंदी  |  Instagram English · తెలుగు · हिंदी