Ziwen Pan — accepted AI-conference papers

2026

Algorithmic Primitives and Compositional Geometry of Reasoning in Language Models

Samuel Lippl, Thomas McGee, Kimberly Lopez, Ziwen Pan, Pierce Zhang, Salma Ziadi +2

ICML 2026poster

How do latent and inference time computations enable large language models (LLMs) to solve multi-step reasoning? We introduce a framework for tracing and steering algorithmic primitives that underlie model reasoning. Our approach links reasoning traces to internal activations and evaluates algorithm…

Cited by 0 Source ↗Scholar ↗

2025

Causal Representation Learning from Multimodal Clinical Records under Non-Random Modality Missingness

Zihan Liang, Ziwen Pan, Ruoxuan Xiong

EMNLP 2025

Clinical notes contain rich patient information, such as diagnoses or medications, making them valuable for patient representation learning. Recent advances in large language models have further improved the ability to extract meaningful representations from clinical texts. However, clinical notes a

Cited by 0 Source ↗PDF ↗Scholar ↗