Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts for Open-Domain QA?

While auxiliary information has become a key to enhance Large Language Models (LLMs), relatively little is known about how LLMs merge these contexts, specifically generated and retrieved. To study this, we formulate a systematic framework to identify whether LLMs' responses, derived from the integration of generated and retrieved contexts, are attributed to either generated or retrieved contexts. To achieve this, we construct datasets with conflicting contexts, where each question is paired with both generated and retrieved contexts, yet only one of them contains the correct answer. Our experiments reveal a significant bias in LLMs (GPT-4/3.5 and Llama2) towards generated contexts, even when they provide incorrect information. We further identify two key factors contributing to this bias: i) contexts generated by LLMs typically show greater similarity to the questions, increasing their likelihood of selection; ii) the segmentation process used in retrieved contexts disrupts their completeness, thereby hindering their full utilization in LLMs. Our analysis enhances the understanding of how LLMs merge diverse contexts, offering valuable insights for advancing current augmentation methods for LLMs.