Saved in:
Bibliographic Details
Main Authors: Yadav, Anushka, Nalawade, Isha, Pillarichety, Srujana, Babu, Yashwanth, Ghosh, Reshmi, Basu, Samyadeep, Zhao, Wenlong, Nasaeh, Ali, Balasubramanian, Sriram, Srinivasan, Soundararajan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.04699
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915432556920832
author Yadav, Anushka
Nalawade, Isha
Pillarichety, Srujana
Babu, Yashwanth
Ghosh, Reshmi
Basu, Samyadeep
Zhao, Wenlong
Nasaeh, Ali
Balasubramanian, Sriram
Srinivasan, Soundararajan
author_facet Yadav, Anushka
Nalawade, Isha
Pillarichety, Srujana
Babu, Yashwanth
Ghosh, Reshmi
Basu, Samyadeep
Zhao, Wenlong
Nasaeh, Ali
Balasubramanian, Sriram
Srinivasan, Soundararajan
contents The emergence of reasoning models and their integration into practical AI chat bots has led to breakthroughs in solving advanced math, deep search, and extractive question answering problems that requires a complex and multi-step thought process. Yet, a complete understanding of why these models hallucinate more than general purpose language models is missing. In this investigative study, we systematicallyexplore reasoning failures of contemporary language models on multi-hop question answering tasks. We introduce a novel, nuanced error categorization framework that examines failures across three critical dimensions: the diversity and uniqueness of source documents involved ("hops"), completeness in capturing relevant information ("coverage"), and cognitive inefficiency ("overthinking"). Through rigorous hu-man annotation, supported by complementary automated metrics, our exploration uncovers intricate error patterns often hidden by accuracy-centric evaluations. This investigative approach provides deeper insights into the cognitive limitations of current models and offers actionable guidance toward enhancing reasoning fidelity, transparency, and robustness in future language modeling efforts.
format Preprint
id arxiv_https___arxiv_org_abs_2508_04699
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hop, Skip, and Overthink: Diagnosing Why Reasoning Models Fumble during Multi-Hop Analysis
Yadav, Anushka
Nalawade, Isha
Pillarichety, Srujana
Babu, Yashwanth
Ghosh, Reshmi
Basu, Samyadeep
Zhao, Wenlong
Nasaeh, Ali
Balasubramanian, Sriram
Srinivasan, Soundararajan
Computation and Language
Artificial Intelligence
The emergence of reasoning models and their integration into practical AI chat bots has led to breakthroughs in solving advanced math, deep search, and extractive question answering problems that requires a complex and multi-step thought process. Yet, a complete understanding of why these models hallucinate more than general purpose language models is missing. In this investigative study, we systematicallyexplore reasoning failures of contemporary language models on multi-hop question answering tasks. We introduce a novel, nuanced error categorization framework that examines failures across three critical dimensions: the diversity and uniqueness of source documents involved ("hops"), completeness in capturing relevant information ("coverage"), and cognitive inefficiency ("overthinking"). Through rigorous hu-man annotation, supported by complementary automated metrics, our exploration uncovers intricate error patterns often hidden by accuracy-centric evaluations. This investigative approach provides deeper insights into the cognitive limitations of current models and offers actionable guidance toward enhancing reasoning fidelity, transparency, and robustness in future language modeling efforts.
title Hop, Skip, and Overthink: Diagnosing Why Reasoning Models Fumble during Multi-Hop Analysis
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2508.04699