no code implementations • LTEDI (ACL) 2022 • Yoon A Park, Frank Rudzicz
Existing studies have investigated the tendency of autoregressive language models to generate contexts that exhibit undesired biases and toxicity.
Text Generation