Saved in:
Bibliographic Details
Main Authors: Diallo, Aissatou, Bikakis, Antonis, Dickens, Luke, Hunter, Anthony, Miller, Rob
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.11348
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916652678905856
author Diallo, Aissatou
Bikakis, Antonis
Dickens, Luke
Hunter, Anthony
Miller, Rob
author_facet Diallo, Aissatou
Bikakis, Antonis
Dickens, Luke
Hunter, Anthony
Miller, Rob
contents An interesting class of commonsense reasoning problems arises when people are faced with natural disasters. To investigate this topic, we present \textsf{RESPONSE}, a human-curated dataset containing 1789 annotated instances featuring 6037 sets of questions designed to assess LLMs' commonsense reasoning in disaster situations across different time frames. The dataset includes problem descriptions, missing resources, time-sensitive solutions, and their justifications, with a subset validated by environmental engineers. Through both automatic metrics and human evaluation, we compare LLM-generated recommendations against human responses. Our findings show that even state-of-the-art models like GPT-4 achieve only 37\% human-evaluated correctness for immediate response actions, highlighting significant room for improvement in LLMs' ability for commonsense reasoning in crises.
format Preprint
id arxiv_https___arxiv_org_abs_2503_11348
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RESPONSE: Benchmarking the Ability of Language Models to Undertake Commonsense Reasoning in Crisis Situation
Diallo, Aissatou
Bikakis, Antonis
Dickens, Luke
Hunter, Anthony
Miller, Rob
Computation and Language
An interesting class of commonsense reasoning problems arises when people are faced with natural disasters. To investigate this topic, we present \textsf{RESPONSE}, a human-curated dataset containing 1789 annotated instances featuring 6037 sets of questions designed to assess LLMs' commonsense reasoning in disaster situations across different time frames. The dataset includes problem descriptions, missing resources, time-sensitive solutions, and their justifications, with a subset validated by environmental engineers. Through both automatic metrics and human evaluation, we compare LLM-generated recommendations against human responses. Our findings show that even state-of-the-art models like GPT-4 achieve only 37\% human-evaluated correctness for immediate response actions, highlighting significant room for improvement in LLMs' ability for commonsense reasoning in crises.
title RESPONSE: Benchmarking the Ability of Language Models to Undertake Commonsense Reasoning in Crisis Situation
topic Computation and Language
url https://arxiv.org/abs/2503.11348