paper-with-me

Papers

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

2025-02-24 · Jie Zeng, Qianyu He, Qingyu Ren, Jiaqing Liang, Yanghua Xiao, Weikang Zhou, Zeye Sun, Fei Yu

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.

📄 PDF Abstract BibTeX arXiv:2502.17204

Code (1)

meowpass/pbif 공식 구현 pytorch

Tasks

Instruction FollowingPosition

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음

Similar Papers 제목 키워드 기반

Local Structure Matters Most: Perturbation Study in NLU

2021-07-29 · Findings (ACL) 2022 5 · Louis Clouatre, Prasanna Parthasarathi, Amal Zouaq, Sarath Chandar

Recent research analyzing the sensitivity of natural language understanding models to word-order perturbations has shown that neural models are surprisingly insensitive to the order of words. In this paper, we investigat…

Natural Language UnderstandingPosition

Local Structure Matters Most: Perturbation Study in NLU

2021-11-16 · ACL ARR November 2021 11 · Anonymous

Recent research analyzing the sensitivity of natural language understanding models to word-order perturbations has shown that neural models are surprisingly insensitive to the order of words. In this paper, we investigat…

Natural Language UnderstandingPosition

Translating Natural Language to SQL using Pointer-Generator Networks and How Decoding Order Matters

2018-11-13 · Denis Lukovnikov, Nilesh Chakraborty, Jens Lehmann, Asja Fischer

Translating natural language to SQL queries for table-based question answering is a challenging problem and has received significant attention from the research community. In this work, we extend a pointer-generator and …

Question AnsweringSemantic ParsingTable-based Question Answering

Attention Basin: Why Contextual Position Matters in Large Language Models

2025-08-07 · Zihao Yi, Delong Zeng, Zhenqing Ling, Haohao Luo 외 arxiv

The performance of Large Language Models (LLMs) is significantly sensitive to the contextual position of information in the input. To investigate the mechanism behind this positional bias, our extensive experiments revea…

Order Matters: Exploring Order Sensitivity in Multimodal Large Language Models

2024-10-22 · Zhijie Tan, Xu Chu, Weiping Li, Tong Mo

Multimodal Large Language Models (MLLMs) utilize multimodal contexts consisting of text, images, or videos to solve various multimodal tasks. However, we find that changing the order of multimodal input can cause the mod…

In-Context LearningQuestion AnsweringSensitivityVisual Question Answering