始める無料で始める

Quiz Question 3

You want to embed the concept of safety into a pre-trained LLM by fine-tuning it. You have a dataset of prompts and pairs of possible answers, along with labels created by human safety evaluators indicating their preference for one answer over the other. Which technique is the most suitable for fine-tuning the LLM in this scenario?

この演習はコースの一部です

Responsible AI for Developers: Privacy & Safety

コースを見る

実践的なインタラクティブ演習

理論を実践に変える、インタラクティブな演習のひとつをお試しください

演習を開始する