Reducing the Probability of Undesirable Outputs in Language Models Using Probabilistic Inference

Devs

Reducing the Probability of Undesirable Outputs in Language Models Using Probabilistic Inference | Read Paper on Bytez