# Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models

Source: [Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models](<https://www.sota2.com/research/paper/69705deae25ad2fa9d1ed7d8>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-23T05:00:12.895Z

Authors: Hyunjong Ok, Jaeho Lee

Published: 2026-01-20T16:54:22.000Z

## 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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2601.14152>)

[PDF](<https://arxiv.org/pdf/2601.14152.pdf>)
