In a not-yet-peer-reviewed paper, a team of researchers from Northeastern University and the Massachusetts Institute of Technology suggests that large language models (LLM) might be able to learn from ...