CITECONTROL

LLMs show spontaneous persuasion, improve RAG, and detect neologisms

Researchers have developed a pipeline to automatically detect neologisms, or new words, by combining rule-based filtering with LLM classification on a large dataset of Reddit posts. Another study explores "spontaneous p…