About Me

I am a third year Ph.D. candidate advised by Arjun Guha at the Programming Research Laboratory of Northeastern University. My interests lie at the intersection of NLP, programming languages and systems. Specifically, I focus on understanding how to use CodeLLMs safely in software applications, as well as training CodeLLMs for novel applications like low-resource programming languages.

I base all my research on insights from mechanistic interpretability, or the view that model capabilities arise from specific model components.

News

10/22: We released our latest preprint Substance Beats Style studying the interactions between novice programmers and LLMs.
10/15: Looking forward to giving a talk at IROS 2024 in Abu Dhabi.
10/01: MultiPL-T will appear at OOPSLA 2024!

Publications

Lucchetti, F., Wu, Z., Feldman, M. Q., Guha, A., & Anderson, C. J.,(2024). Substance Beats Style: Why Beginning Students Fail to Code with LLMs. Preprint. Paper

Lucchetti, F., & Guha, A. (2024). Understanding How CodeLLMs (Mis)Predict Types with Activation Steering. Preprint. Paper

Hu, Z., Lucchetti, F., Schlesinger, C., Saxena, Y., Freeman, A., Modak, S., ... & Biswas, J. (2024). Deploying and Evaluating LLMs to Program Service Mobile Robots. IEEE Robotics and Automation Letters.

Paper

Demo

Cassano, F., Gouwar, J., Lucchetti, F., Schlesinger, C., Freeman, A., Anderson, C. J., ... & Guha, A. (2024). Knowledge transfer from high-resource to low-resource programming languages for code llms. Proceedings of the ACM on Programming Languages, 8(OOPSLA2), 677-708. Paper