RLHF Rater
Evaluated AI responses for instruction following, helpfulness, safety, and tone. Read and researched answers from AI, applied rating and reasoning for each rating, including notes on how to make the response more helpful, sometimes correcting and rewriting responses