Dyah Adila

UW-Madison CS Dept. adila@wisc.edu.

Profile Picture

Hello! I am a PhD student advised by Fred Sala in the Sprocket Lab.

I’ve been fortunate to intern at Google Research, collaborating with Hanna Mazzawi, Benoit Dherin, and Xavi Gonzalvo as well as at AWS AI Labs.

Research Interest. My work focuses on:

  • Developing data and compute-efficient algorithms for adapting LLMs to specific tasks and preferences
  • Building robust and reliable models that generalize across diverse scenarios
  • Exploring LLM latent representations to uncover emergent patterns and deeper insights into their internal mechanisms

:star2: I’m currently on the job market for industry roles starting in 2026. Always happy to chat about research scientist/engineer opportunities :)

See my CV

news

Mar 1, 2025 Check out our preprint on data-and compute efficient LLM alignment!
Jan 20, 2025 Our paper on ultra-efficient way to personalize language models is acepted at NAACL’25 Findings!

selected publications

  1. Preprint
    Alignment, Simplified: Steering LLMs with Self-Generated Preferences
    Adila, Dyah, Shin, Changho, Zhang, Yijing, and Sala, Frederic
    In Preprint Jan 2025
    PDF
  2. ICML 2024
    Discovering Bias in Latent Space: An Unsupervised Debiasing Approach
    Adila, Dyah, Zhang, Shuai, Han, Boran, and Wang, Bernie
    In ICML 2024 Jul 2024
    PDF
  3. ICLR 2024
    Zero-Shot Robustification of Zero-Shot Models
    Adila, Dyah, Shin, Changho, Cai, Lingrong, and Sala, Frederic
    In ICLR 2024 May 2024
    PDF
  4. UAI
    Shoring Up the Foundations: Fusing Model Embeddings and Weak Supervision
    Chen, Mayee F, Fu, Daniel Y, Adila, Dyah, Zhang, Michael, Sala, Frederic, Fatahalian, Kayvon, and Ré, Christopher
    In Proceedings of the conference on Uncertainty in artificial intelligence Aug 2022
    PDF
  5. NeurIPS
    Understanding Out-of-distribution: A Perspective of Data Dynamics
    Adila, Dyah, and Kang, Dongyeop
    In ICBINB@NeurIPS 2021 Nov 2021
    PDF