, where $e_{doc}$ and $e_{query}$ are task-specific document, query descriptions respectively, $e_{prompt}$ is a task-specific prompt/instruction for ChatGPT/GPT-4, and $d$ is a new document, for which LLM will generate a query.
, where $e_{doc}$ and $e_{query}$ are task-specific document, query descriptions respectively, $e_{prompt}$ is a task-specific prompt/instruction for ChatGPT/GPT-4, and $d$ is a new document, for which LLM will generate a query.