AI expert annotator
The project involved evaluating AI response of a user’s prompt request based on scales like Harmlessness, Honesty, and Helpfulness. Conclusions are made on the score and the general preferred response from the responses on those three scales.