Varun R. Embar

Eliciting In-context Retrieval and Reasoning for Long-context Large Language Models

Recent advancements in long-context language models (LCLMs) promise to transform Retrieval-Augmented Generation (RAG) by simplifying pipelines. With their expanded context windows, LCLMs can process entire knowledge bases and perform retrieval and reasoning directly – a capability we define as In-Co…