Watch this video on YouTube
A recommendation engine is being improved using RLHF to better reflect subjective user experiences. Which action should the development team take before training the reinforcement component?