Activation Steering for Conditional Molecular Generation
AI4Mat Workshop @ NeurIPS 2025, 2025
Using concept bottleneck models and activation steering, we enable conditional molecular generation by directly manipulating internal LLM representations.
Publications
You can also find my articles on my Google Scholar profile.
Conference & Workshop Papers
Multilingual Assessment of Stereotypes in Large Language Models
NAACL 2025, 2025
We build and assess multilingual stereotypes across different LLMs.
Atypicality-Aware Calibration of LLMs for Medical QA
Findings of EMNLP 2024, 2024
We propose a novel method for eliciting LLM confidence in Medical QA by leveraging insights from medical atypical presentations.
Preprints
Calibrating Large Language Models with Concept Activation Vectors for Medical QA
Preprint, 2024
We propose a novel framework for calibrating LLM uncertainty through Concept Activation Vectors, improving safety and calibration in high-stakes medical decision making.