DeepThink: Aligning Language Models with Domain-Specific User Intents
This work addresses the challenge of aligning language models with domain-specific user intents, particularly in advertising, but is incremental as it builds on existing fine-tuning and RAG methods.
The paper tackles the problem of synthesized instructions deviating from real user questions in domain-specific QA tasks by proposing DeepThink, a framework that generates high-quality instructions through seed questions, simulated conversations, and document retrieval, resulting in an average performance improvement of 7.92% compared to a GPT-4-turbo+RAG-based assistant on a real user test set in the advertising domain.
Supervised fine-tuning with synthesized instructions has been a common practice for adapting LLMs to domain-specific QA tasks. However, the synthesized instructions deviate from real user questions and expected answers. This study proposes a novel framework called DeepThink to generate high-quality instructions. DeepThink first generates a few seed questions to mimic actual user questions, simulates conversations to uncover the hidden user needs, and refines the answer by conversational contexts and the retrieved documents for more comprehensive answers. Experiments demonstrate that DeepThink achieves an average performance improvement of 7.92% compared to a GPT-4-turbo+RAG-based assistant on the real user test set in the advertising domain across dimensions such as relevance, completeness, clarity, accuracy, and actionability.