Order Matters: Investigate the Position Bias in Multi-constraint Instruction Following

TL;DR

This paper investigates how the order of multiple constraints affects large language models' performance in instruction following, revealing a preference for a 'hard-to-easy' order and analyzing the underlying attention mechanisms.

Contribution

It introduces a novel Difficulty Distribution Index (CDDI) to quantify constraint difficulty and systematically studies position bias in multi-constraint instruction tasks.

Findings

LLMs perform better with a 'hard-to-easy' constraint order

The position bias persists across different architectures and sizes

Attention mechanisms correlate with constraint order effects

Abstract

Real-world instructions with multiple constraints pose a significant challenge to existing large language models (LLMs). An observation is that the LLMs exhibit dramatic performance fluctuation when disturbing the order of the incorporated constraints. Yet, none of the existing works has systematically investigated this position bias problem in the field of multi-constraint instruction following. To bridge this gap, we design a probing task where we quantitatively measure the difficulty distribution of the constraints by a novel Difficulty Distribution Index (CDDI). Through the experimental results, we find that LLMs are more performant when presented with the constraints in a ``hard-to-easy'' order. This preference can be generalized to LLMs with different architecture or different sizes of parameters. Additionally, we conduct an explanation study, providing an intuitive insight into the correlation between the LLM's attention and constraint orders. Our code and dataset are publicly available at https://github.com/meowpass/PBIF.