Research map / Large language models / Alignment and preferences
QueryAligner: Customizing User Query to Match LLMs Preferences for Better Intent Recognition
Yunlong Ma, Bo Wang, Yihong Tang, Zifei Yu, Chenyun Xue, Gaoke Zhang, Yuexian Hou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add approximation matching (similarity filtering of reverse-engineered pairs) (ablation) | QueryAligner · Natural Questions | accuracy: 51.01 -> 54.83 | Improved |
| add supervised learning stage (ablation) | QueryAligner · Natural Questions | accuracy: 33.52 -> 54.83 | Improved |
| add reinforcement learning stage (ablation) | QueryAligner · Natural Questions | accuracy: 51.52 -> 54.83 | Improved |
| add CoT to QueryAligner (ablation) | QueryAligner (gpt-4o-mini) · Natural Questions | accuracy: 53.11 -> 54.83 | Improved |
| use QueryAligner rewriting vs plain Naive RAG | gpt-4o-mini RAG · Natural Questions | accuracy: 50.69 -> 54.83 | Improved |
| use QueryAligner rewriting vs plain Naive RAG | Llama-2-13b RAG · HotpotQA | accuracy: 67.90 -> 74.31 | Improved |
| use QueryAligner vs Naive RAG | gpt-4o-mini RAG · PubMedQA | accuracy: 93.82 -> 95.44 | Improved |
| prepend QueryAligner to existing methods (CoT-SC, RaR, RAFT, Self-RAG, Naive RAG) | existing RAG methods · Natural Questions | accuracy: +2.12 average | Improved |
| train rewriter on one dataset, fine-tune 10% on target | QueryAligner · cross-dataset (HotpotQA to PubMedQA) | accuracy vs Naive RAG: -0.27 | Hurt |
| train rewriter on another dataset (transfer) | QueryAligner · cross-dataset (other bases to NQ/HotpotQA) | accuracy vs Naive RAG: +0.36 to +4.44 | Improved |
| QueryAligner vs RaR on gpt-4o-mini PubMedQA EM | QueryAligner · PubMedQA | EM: 92.59 -> 93.81 | Improved |
Closest papers by meaning
- LLM Alignment as Retriever Optimization: An Information Retrieval Perspective (ICML 2025 · similarity 0.90)
- The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning (ICLR 2024 · similarity 0.89)
- LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations (NeurIPS 2026 · similarity 0.88)
- Reward-Augmented Data Enhances Direct Preference Alignment of LLMs (ICML 2025 · similarity 0.88)
- Panacea: Pareto Alignment via Preference Adaptation for LLMs (NeurIPS 2024 · similarity 0.88)
- Trustworthy Alignment of Retrieval-Augmented Large Language Models via Reinforcement Learning (ICML 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
