AI.info

Jan Dubiński

Explore Jan Dubiński on AI.info.

• Reference-Based Bias Detection in LLMs via Relative Representations of Hidden States
• On Stealing Graph Neural Network Models