Alessandro Stolfo

prof_pic.jpeg

Hi, I am a research scientist at Google DeepMind working on Gemini post-training.

I obtained my PhD in the Institute for Machine Learning at ETH Zürich, where I was advised by Prof. Mrinmaya Sachan, and co-advised by Prof. Yonatan Belinkov (Technion). My PhD research revolved around LLM interpretability and was supported by the CYD Doctoral Fellowship.

During my PhD, I interned with the AI Frontiers group at Microsoft Research in Redmond, WA, where I had the opportunity to collaborate with Besmira Nushi and Eric Horvitz. Previously, I interned with the Machine Learning Research Group at Oracle Labs in Burlington, MA, working with Ari Kobren.

news

Jul 24, 2025 Gave a talk at NEC Labs EU about our recent work on LLM steering. Check it out on YouTube.
May 28, 2024 Interning in the AI Frontiers group at Microsoft Research in Redmond, WA.
Nov 22, 2023 Attending the ML Alignment & Theory Scholars (MATS) Program, mentored by Neel Nanda.
Jul 17, 2023 Started my internship in the ML Research Group at Oracle Labs in Burlington, MA.
Apr 22, 2022 Answered a couple of questions for this EPFL News article. Check it out!

selected publications

  1. NeurIPS 2025
    Dense SAE Latents Are Features, Not Bugs
    X. Sun*, A. Stolfo*, J. Engels, B. Wu, S. Rajamanoharan, M. Sachan, and M. Tegmark
    OpenReview arXiv
  2. ICLR 2025
    Improving Instruction-Following in Language Models through Activation Steering
    A. Stolfo, V. Balachandran, S. Yousefi, E. Horvitz, and B. Nushi
    OpenReview arXiv
  3. NeurIPS 2024
    Confidence Regulation Neurons in Language Models
    A. Stolfo*, B. Wu*, W. Gurnee, Y. Belinkov, X. Song, M. Sachan, and N. Nanda
    OpenReview arXiv
  4. ICML 2024
    Do Language Models Exhibit the Same Cognitive Biases in Problem Solving as Human Learners?
    A. Opedal*, A. Stolfo*, H. Shirakami, Y. Jiao, R. Cotterell, B. Schölkopf, A. Saparov, and M. Sachan
    OpenReview arXiv
  5. EMNLP 2023
    A Mechanistic Interpretation of Arithmetic Reasoning in Language Models using Causal Mediation Analysis
    A. Stolfo, Y. Belinkov, and M. Sachan
    ACL Anthology arXiv
  6. ACL 2023
    A Causal Framework to Quantify the Robustness of Mathematical Reasoning with Language Models
    A. Stolfo*, Z. Jin*, K. Shridhar, B. Schölkopf, and M. Sachan
    ACL Anthology arXiv