# Semantically Aligned Question and Code Generation for Automated Insight Generation

arXiv: https://arxiv.org/abs/2405.01556 · HTML: https://mlsanity.com/paper/2405.01556/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 19 references that are in the map.
- First posted on arXiv 21 Mar 2024.

## Builds on (most cited first, 9 of 9)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generation](https://mlsanity.com/paper/2109.00859/) (2021)
- [Text and Code Embeddings by Contrastive Pre-Training](https://mlsanity.com/paper/2201.10005/) (2022)
- [Large Language Models Meet NL2Code: A Survey](https://mlsanity.com/paper/2212.09420/) (2022)
- [Jigsaw: Large Language Models meet Program Synthesis](https://mlsanity.com/paper/2112.02969/) (2021)
- [When Deep Learning Met Code Search](https://mlsanity.com/paper/1905.03813/) (2019)
- [Automatic Generation of Socratic Subquestions for Teaching Math Word Problems](https://mlsanity.com/paper/2211.12835/) (2022)
- [Open-WikiTable: Dataset for Open Domain Question Answering with Complex Reasoning over Table](https://mlsanity.com/paper/2305.07288/) (2023)
- [Lux: Always-on Visualization Recommendations for Exploratory Dataframe Workflows](https://mlsanity.com/paper/2105.00121/) (2021)

## Built on by (most cited first, 3 of 3)

- [Diffusion is a code repair operator and generator](https://mlsanity.com/paper/2508.11110/) (2025)
- [MDSF: Context-Aware Multi-Dimensional Data Storytelling Framework based on Large language Model](https://mlsanity.com/paper/2501.01014/) (2025)
- [Tabularis Formatus: Predictive Formatting for Tables](https://mlsanity.com/paper/2508.11121/) (2025)

## Abstract (arXiv)

Automated insight generation is a common tactic for helping knowledge workers, such as data scientists, to quickly understand the potential value of new and unfamiliar data. Unfortunately, automated insights produced by large-language models can generate code that does not correctly correspond (or align) to the insight. In this paper, we leverage the semantic knowledge of large language models to generate targeted and insightful questions about data and the corresponding code to answer those questions. Then through an empirical study on data from Open-WikiTable, we show that embeddings can be effectively used for filtering out semantically unaligned pairs of question and code. Additionally, we found that generating questions and code together yields more diverse questions.