Whispering experts: mitigating toxicity in pre-trained language models by attenuating expert neurons by Technical Terrence Team 07/21/2024 0 A major problem with large language models (LLMs) is their unintended ability to generate toxic language. In this work, we ...
Google AI Introduces Visually Rich Document Understanding (VRDU): A Dataset for Better Tracking of Document Understanding Task Progress 08/12/2023