Atom-Searcher: Enhancing Agentic Deep Research via Fine-Grained Atomic Thought Reward

Yong Deng, Guoqing Wang, Zhenzhe Ying +12

Large language models (LLMs) exhibit remarkable problem-solving abilities, but struggle with complex tasks due to static internal knowledge. Retrieval-Augmented Generation (RAG) en…