1 paper · 1 filter
Walter Laurito, Sharan Maiya, Grégoire Dhimoïla +3
The deployment of language models brings challenges in generating reliable information, especially when these models are fine-tuned using human preferences. To extract encoded know…