PromptTailor: Multi-turn Intent-Aligned Prompt Synthesis for Lightweight LLMs

2025-11-20

AI Summary

Lightweight language models exhibit high sensitivity to prompt quality in open-ended generation, yet non-expert users struggle to consistently craft high-quality prompts, and existing prompt optimization methods often deviate from user intent. To address this, we propose an intent-aligned multi-round prompt synthesis framework: (1) We distill 12,300 cross-domain dialogues from multiple strong LLMs to train a lightweight LoRA adapter that preserves user intent while enabling domain-aware prompt expansion; (2) We adopt quantized Llama3-8B as the base model for efficient edge deployment. Experiments demonstrate that our method outperforms chain-of-thought prompting across multiple models and benchmarks, matches or exceeds state-of-the-art prompt optimization approaches, and requires only three model calls—reducing inference cost by 67% versus baselines—while significantly improving user preference rates.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Lightweight language models remain attractive for on-device and privacy-sensitive applications, but their responses are highly sensitive to prompt quality. For open-ended generation, non-expert users often lack the knowledge or time to consistently craft high-quality prompts, leading them to rely on prompt optimization tools. However, a key challenge is ensuring the optimized prompts genuinely align with users'original intents and preferences. We introduce PromptTailor, a system for controllable prompt generation for open-ended text that improves model output quality by intent-aligned prompt synthesis. PromptTailor expands minimal user instructions into rich, domain-aware prompts while preserving the user's stated preferences. The system is a quantized Llama3-8B model fine-tuned with a lightweight LoRA adapter on 12,300 prompt-refinement dialogues spanning 41 everyday domains, distilled from three stronger LLMs. The adapter attaches to any Llama3-8B base, enabling edge deployment. In human and LLM-judge evaluations across multiple target models and optimization baselines, PromptTailor yields higher preference rates than chain-of-thought prompting and matches or surpasses state-of-the-art prompt optimization methods while requiring fewer model calls (e.g., 3 vs. 9). These results show that a compact student, guided by powerful teachers, can learn effective prompt-generation strategies that enhance response quality while maintaining alignment with user intent.

Problem

• Optimizes prompts for lightweight LLMs to improve response quality
• Ensures generated prompts align with user intent and preferences
• Enables efficient prompt synthesis for on-device deployment

Innovation

• Fine-tuned Llama3-8B with LoRA for prompt synthesis
• Expands user instructions into domain-aware prompts
• Reduces model calls while matching state-of-the-art performance