New AI Method Makes Small Language Models Better at Table Reasoning
Researchers have developed a method called RSAT that helps small language models explain their reasoning when answering questions about tables. This makes it easier for users to verify the accuracy of the AI's answers.