arXiv:2601.14152v2 Announce Type: replace-cross
Abstract: Large language models exhibit surprising sensitivity to the structure of the prompt, but the mechanisms underlying this sensitivity remain poorly understood. In this work, we conduct an in-depth investigation on a striking case: in multiple-choice question answering, placing context before the questions and options (CQO) outperforms the reverse order (QOC) by over 14%p, consistently over a wide range of models and datasets. Through systematic architectural analysis, we identify causal attention as the core mechanism: in QOC prompts, the causal mask prevents option tokens from attending to context, creating an information bottleneck where context becomes invisible to options.
Coordinated Temporal Dynamics of Glucocorticoid Receptor Binding and Chromatin Landscape Drive Transcriptional Regulation
Glucocorticoid receptor (GR) signaling elicits diverse transcriptional responses through dynamic and context-dependent interactions with chromatin. Here, we define a temporally resolved and mechanistically integrated framework