Weak-to-Strong Prompts with Lightweight-to-Powerful LLMs for High-Accuracy, Low-Cost, and Explainable Data Transformation
Summary: MegaTran turns underspecified data-transformation requests into structured prompts with a lightweight LLM, then uses a powerful LLM for code generation. Checklist-based reflection and LazyRAG improve correctness, explainability, and cost efficiency, yielding 2.2–26.1% accuracy gains. (summarized by gpt-5.6-luna on Jul 24 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Changlun Li (Hong Kong University of Science and Technology)
- 2. Chenyu Yang (Hong Kong University of Science and Technology)
- 3. Yuyu Luo (Hong Kong University of Science and Technology)
- 4. Ju Fan (Renmin University of China)
- 5. Nan Tang (Hong Kong University of Science and Technology)
BibTeX Citation
@article{li_vldb25,
title = {{Weak-to-Strong Prompts with Lightweight-to-Powerful LLMs for High-Accuracy, Low-Cost, and Explainable Data Transformation}},
author = {Li, Changlun and Yang, Chenyu and Luo, Yuyu and Fan, Ju and Tang, Nan},
journal = {PVLDB},
series = {{VLDB} '25},
volume = {18},
number = {8},
pages = {2371--2384},
doi = {10.14778/3742728.3742734},
url = {https://doi.org/10.14778/3742728.3742734},
year = {2025}
}Incoming Citations (Sorted by Pagerank)
Showing 1 of 1 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 10,587 | LEAD: Iterative Data Selection for Efficient LLM Instruction Tuning | 2026 | VLDB | 5.093636e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 20 of 20 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 10,856 | Optimized Batch Prompting for Cost-effective LLMs | 2025 | VLDB |
| 2 | 6,545 | DTT: An Example-Driven Tabular Transformer for Joinability by Leveraging Large Language Models | 2024 | SIGMOD |
| 3 | 4,886 | Chat2Data: An Interactive Data Analysis System with RAG, Vector Databases and LLMs | 2024 | VLDB |
| 4 | 8,140 | Automated Data Visualization from Natural Language via Large Language Models: An Exploratory Study | 2024 | SIGMOD |
| 5 | 8,829 | Cut Costs, Not Accuracy: LLM-Powered Data Processing with Guarantees | 2026 | SIGMOD |
| 6 | 713 | Language Models Enable Simple Systems for Generating Structured Views of Heterogeneous Data Lakes | 2024 | VLDB |
| 7 | 10,614 | LLM-AutoDP: Automatic Data Processing via LLM Agents for Model Fine-tuning | 2026 | VLDB |
| 8 | 4,050 | Revisiting Prompt Engineering via Declarative Crowdsourcing | 2024 | CIDR |
| 9 | 9,505 | TabulaX: Leveraging Large Language Models for Multi-Class Table Transformations | 2025 | VLDB |
| 10 | 6,101 | LLM for Data Management | 2024 | VLDB |