paper-with-me

Papers

Understanding and Patching Compositional Reasoning in LLMs

2024-02-22 · Zhaoyi Li, Gangwei Jiang, Hong Xie, Linqi Song, Defu Lian, Ying WEI

LLMs have marked a revolutonary shift, yet they falter when faced with compositional reasoning tasks. Our research embarks on a quest to uncover the root causes of compositional reasoning failures of LLMs, uncovering that most of them stem from the improperly generated or leveraged implicit reasoning results. Inspired by our empirical findings, we resort to Logit Lens and an intervention experiment to dissect the inner hidden states of LLMs. This deep dive reveals that implicit reasoning results indeed surface within middle layers and play a causative role in shaping the final explicit reasoning results. Our exploration further locates multi-head self-attention (MHSA) modules within these layers, which emerge as the linchpins in accurate generation and leveraing of implicit reasoning results. Grounded on the above findings, we develop CREME, a lightweight method to patch errors in compositional reasoning via editing the located MHSA modules. Our empirical evidence stands testament to CREME's effectiveness, paving the way for autonomously and continuously enhancing compositional reasoning capabilities in language models.

📄 PDF Abstract BibTeX arXiv:2402.14328

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Benchmarking and Understanding Compositional Relational Reasoning of LLMs

2024-12-17 · Ruikang Ni, Da Xiao, Qingye Meng, Xiangyu Li 외

Compositional relational reasoning (CRR) is a hallmark of human intelligence, but we lack a clear understanding of whether and how existing transformer large language models (LLMs) can solve CRR tasks. To enable systemat…

BenchmarkingRelational Reasoning

COCO-Tree: Compositional Hierarchical Concept Trees for Enhanced Reasoning in Vision Language Models

2025-10-13 · Sanchit Sinha, Guangzhi Xiong, Aidong Zhang arxiv

Compositional reasoning remains a persistent weakness of modern vision language models (VLMs): they often falter when a task hinges on understanding how multiple objects, attributes, and relations interact within an imag…

Dissecting Persona-Driven Reasoning in Language Models via Activation Patching

2025-07-28 · Ansh Poonia, Maeghal Jain arxiv

Large language models (LLMs) exhibit remarkable versatility in adopting diverse personas. In this study, we examine how assigning a persona influences a model's reasoning on an objective task. Using activation patching, …

STEP: Enhancing Video-LLMs' Compositional Reasoning by Spatio-Temporal Graph-guided Self-Training

2024-11-29 · CVPR 2025 1 · Haiyi Qiu, Minghe Gao, Long Qian, Kaihang Pan 외

Video Large Language Models (Video-LLMs) have recently shown strong performance in basic video understanding tasks, such as captioning and coarse-grained question answering, but struggle with compositional reasoning that…

Question AnsweringVideo Understanding

Exploring the Compositional Deficiency of Large Language Models in Mathematical Reasoning

2024-05-05 · Jun Zhao, Jingqi Tong, Yurong Mou, Ming Zhang 외

Human cognition exhibits systematic compositionality, the algebraic ability to generate infinite novel combinations from finite learned components, which is the key to understanding and reasoning about complex logic. In …

GSM8KMathMathematical Reasoning