Sophie Ostmeier

Symbal: Detecting Systematic Misalignments in Model-Generated Captions

Attention Head Entropy of LLMs Predicts Answer Correctness

LieRE: Lie Rotational Positional Encodings

TRoVe: Discovering Error-Inducing Static Feature Biases in Temporal Vision-Language Models

LieRE: Generalizing Rotary Position Encodings to Higher Dimensional Inputs

Preference fine-tuning for factuality in chest X-ray interpretation models without human feedback