# A Study on Developer Behaviors for Validating and Repairing LLM-Generated Code Using Eye Tracking and IDE Actions

arXiv: https://arxiv.org/abs/2405.16081 · HTML: https://mlsanity.com/paper/2405.16081/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 May 2024.

## Built on by (most cited first, 12 of 12)

- [AIDev: Studying AI Coding Agents on GitHub](https://mlsanity.com/paper/2602.09185/) (2026)
- [Conversational AI as a Coding Assistant: Understanding Programmers' Interactions with and Expectations from Large Language Models for Coding](https://mlsanity.com/paper/2503.16508/) (2025)
- [Risk Assessment Framework for Code LLMs via Leveraging Internal States](https://mlsanity.com/paper/2504.14640/) (2025)
- [Evaluating and Improving Automated Repository-Level Rust Issue Resolution with LLM-based Agents](https://mlsanity.com/paper/2602.22764/) (2026)
- [Human-AI Experience in Integrated Development Environments: A Systematic Literature Review](https://mlsanity.com/paper/2503.06195/) (2025)
- [Engineering Reliable Coding Agents: Evaluating and Operating the System Around the Model](https://mlsanity.com/paper/2608.13867/) (2026)
- [Cheap Code, Costly Judgment: A Case Study on Governable Agentic Software Engineering](https://mlsanity.com/paper/2607.01087/) (2026)
- [DeepCodeProbe: Towards Understanding What Models Trained on Code Learn](https://mlsanity.com/paper/2407.08890/) (2024)
- [Empowering Autonomous Debugging Agents with Efficient Dynamic Analysis](https://mlsanity.com/paper/2604.24212/) (2026)
- [Cracking CodeWhisperer: Analyzing Developers' Interactions and Patterns During Programming Tasks](https://mlsanity.com/paper/2510.11516/) (2025)
- [Choose Your Own Adventure: Non-Linear AI-Assisted Programming with EvoGraph](https://mlsanity.com/paper/2604.18883/) (2026)
- [Software Delegation Contracts: Measuring Reviewability in AI Coding-Agent Work](https://mlsanity.com/paper/2606.17099/) (2026)

## Abstract (arXiv)

The increasing use of large language model (LLM)-powered code generation tools, such as GitHub Copilot, is transforming software engineering practices. This paper investigates how developers validate and repair code generated by Copilot and examines the impact of code provenance awareness during these processes. We conducted a lab study with 28 participants, who were tasked with validating and repairing Copilot-generated code in three software projects. Participants were randomly divided into two groups: one informed about the provenance of LLM-generated code and the other not. We collected data on IDE interactions, eye-tracking, cognitive workload assessments, and conducted semi-structured interviews. Our results indicate that, without explicit information, developers often fail to identify the LLM origin of the code. Developers generally employ similar validation and repair strategies for LLM-generated code, but exhibit behaviors such as frequent switching between code and comments, different attentional focus, and a tendency to delete and rewrite code. Being aware of the code's provenance led to improved performance, increased search efforts, more frequent Copilot usage, and higher cognitive workload. These findings enhance our understanding of how developers interact with LLM-generated code and carry implications for designing tools that facilitate effective human-LLM collaboration in software development.