Daniel Scalena

prof_pic.jpg
Milan, Italy

Hi! I am Daniel, a (double) first-year PhD student at the 🇮🇹 University of Milano - Bicocca and the 🇳🇱 University of Groningen working on interpretability, fairness and security of generative (and non-generative) Large Language Models. My supervisors are Elisabetta Fersini and Malvina Nissim.

My research focuses on the use of interpretability as a tool to make generative models safer, more reliable and less toxic to extend and improve their real-world applications.

In my spare time I take pictures and echo "from NL import infrastructure" > Milan.py.

news

Oct 02, 2024 📜 Multi-property Steering paper accepted to BlackBoxNLP 2024 (@EMNLP 2024) and 📜 A gentle push funziona benissimo accepted @ CLIC-it conference! 🎉
Jun 26, 2024 📚 New work available on arXiv: Multi-property Steering of Large Language Models with Dynamic Activation Composition
Dec 07, 2023 Presenting my poster and paper at the BlackBoxNLP workshop @EMNLP 2023 in Singapore 🇸🇬
Oct 26, 2023 Graduated! Thesis here 🎓

latest posts

selected publications

  1. A gentle push funziona benissimo: making instructed models in Italian via contrastive activation steering
    Daniel Scalena, Elisabetta Fersini, and Malvina Nissim
    2024
  2. duckSteering.jpeg
    Multi-property Steering of Large Language Models with Dynamic Activation Composition
    Daniel Scalena, Gabriele Sarti, and Malvina Nissim
    2024
  3. LetTheModelRespond.png
    Let the Models Respond: Interpreting Language Model Detoxification Through the Lens of Prompt Dependence
    Daniel Scalena, Gabriele Sarti, Malvina Nissim, and 1 more author
    2023