arXiv: Quantifying Claim Smoothing by LLM Research Agents
Deep research, or any use of large language models (LLMs) to read a large document collection and produce insights, requires synthesizing information from different sources.
The arXiv Language-Model Research page documents the arXiv update.