Chung-Hsiang Lo

Chung-Hsiang Lo, Lu Li, Diji Yang et al.

Large language models (LLMs) conventionally process structured inputs as 1D token sequences. While natural for prose, such linearization may introduce additional representational burden for tasks whose computation depends directly on explicit 2D structure, because row--column alignment and local neighborhoods are no longer directly expressed in the input. We study this setting, which we refer to as serialization friction, on a small diagnostic testbed of synthetic tasks with explicit 2D structure: matrix transpose, Conway's Game of Life, and LU decomposition. To examine this question, we compare a text-only language pathway over serialized inputs with a vision-augmented pathway, built on the same language backbone, that receives the same underlying content rendered in task-faithful 2D layout, yielding a system-level comparison between two end-to-end input pathways. Across the tasks and settings we study, the visual pathway consistently outperforms the textual pathway; the gap often widens at larger dimensions, and error patterns under serialization become increasingly spatially structured. These findings indicate that the relationship between input representation and model performance on such tasks warrants further investigation, and suggest that preserving task-relevant 2D layout is a promising direction for structured 2D tasks.

Xinnan Dai, Kai Guo, Chung-Hsiang Lo et al.

Large Language Models (LLMs) demonstrate remarkable reasoning capabilities, yet the structural mechanisms underlying these abilities remain under explored. In this work, we introduce GraphGhost, a unified framework that represents neuron activations and their signal propagation as graphs, explaining how LLMs capture structural semantics from sequential inputs and generate outputs through structurally consistent mechanisms. This graph-based perspective enables us to employ graph algorithms such as PageRank to characterize the properties of LLMs, revealing both shared and model-specific reasoning behaviors across diverse datasets. We further identify the activated neurons within GraphGhost and evaluate them through structural interventions, showing that edits to key neuron nodes can trigger reasoning collapse, altering both logical flow and semantic understanding. Together, these contributions position GraphGhost as a powerful tool for analyzing, intervening in, and ultimately understanding the structural foundations of reasoning in LLMs.

Xinnan Dai, Chung-Hsiang Lo, Kai Guo et al.

Transformer-based LLMs demonstrate strong performance on graph reasoning tasks, yet their internal mechanisms remain underexplored. To uncover these reasoning process mechanisms in a fundamental and unified view, we set the basic decoder-only transformers and explain them using the circuit-tracer framework. Through this lens, we visualize reasoning traces and identify two core mechanisms in graph reasoning: token merging and structural memorization, which underlie both path reasoning and substructure extraction tasks. We further quantify these behaviors and analyze how they are influenced by graph density and model size. Our study provides a unified interpretability framework for understanding structural reasoning in decoder-only Transformers.