1 paper
Justin Lovelace, Varsha Kishore, Yiwei Chen +1
Current language models demonstrate remarkable proficiency in text generation. However, for many applications it is desirable to control attributes, such as sentiment, or toxicity,…