Common Sense vs. Morality: The Curious Case of Narrative Focus Bias in LLMs

• URL: http://arxiv.org/abs/2603.09434v1
• Date: Tue, 10 Mar 2026 09:47:18 GMT
• Title: Common Sense vs. Morality: The Curious Case of Narrative Focus Bias in LLMs
• Abstract summary: Large Language Models (LLMs) are increasingly deployed across diverse real-world applications and user communities.<n>We uncover a critical limitation of current LLMs -- their tendency to prioritize moral reasoning over commonsense understanding.<n>Our comprehensive analysis underscores the need for enhanced reasoning-aware training to improve the commonsense robustness of large language models.
• Score: 2.4025805669104567
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Large Language Models (LLMs) are increasingly deployed across diverse real-world applications and user communities. As such, it is crucial that these models remain both morally grounded and knowledge-aware. In this work, we uncover a critical limitation of current LLMs -- their tendency to prioritize moral reasoning over commonsense understanding. To investigate this phenomenon, we introduce CoMoral, a novel benchmark dataset containing commonsense contradictions embedded within moral dilemmas. Through extensive evaluation of ten LLMs across different model sizes, we find that existing models consistently struggle to identify such contradictions without prior signal. Furthermore, we observe a pervasive narrative focus bias, wherein LLMs more readily detect commonsense contradictions when they are attributed to a secondary character rather than the primary (narrator) character. Our comprehensive analysis underscores the need for enhanced reasoning-aware training to improve the commonsense robustness of large language models.

Related papers

• The Curious Case of Analogies: Investigating Analogical Reasoning in Large Language Models [22.609819017261632]
  Analogical reasoning is at the core of human cognition, serving as an important foundation for a variety of intellectual activities.<n>While prior work has shown that LLMs can represent task patterns and surface-level concepts, it remains unclear whether these models can encode high-level relational concepts.
  arXiv  Detail & Related papers  (2025-11-25T14:23:58Z)
• MORABLES: A Benchmark for Assessing Abstract Moral Reasoning in LLMs with Fables [50.29407048003165]
  We present MORABLES, a human-verified benchmark built from fables and short stories drawn from historical literature.<n>The main task is structured as multiple-choice questions targeting moral inference, with carefully crafted distractors that challenge models to go beyond shallow, extractive question answering.<n>Our findings show that, while larger models outperform smaller ones, they remain susceptible to adversarial manipulation and often rely on superficial patterns rather than true moral reasoning.
  arXiv  Detail & Related papers  (2025-09-15T19:06:10Z)
• Beyond Prompt-Induced Lies: Investigating LLM Deception on Benign Prompts [79.1081247754018]
  Large Language Models (LLMs) are widely deployed in reasoning, planning, and decision-making tasks.<n>We propose a framework based on Contact Searching Questions(CSQ) to quantify the likelihood of deception.
  arXiv  Detail & Related papers  (2025-08-08T14:46:35Z)
• Preference Leakage: A Contamination Problem in LLM-as-a-judge [69.96778498636071]
  Large Language Models (LLMs) as judges and LLM-based data synthesis have emerged as two fundamental LLM-driven data annotation methods.<n>In this work, we expose preference leakage, a contamination problem in LLM-as-a-judge caused by the relatedness between the synthetic data generators and LLM-based evaluators.
  arXiv  Detail & Related papers  (2025-02-03T17:13:03Z)
• Normative Evaluation of Large Language Models with Everyday Moral Dilemmas [0.0]
  We evaluate large language models (LLMs) on complex, everyday moral dilemmas sourced from the "Am I the Asshole" (AITA) community on Reddit.<n>Our results demonstrate that large language models exhibit distinct patterns of moral judgment, varying substantially from human evaluations on the AITA subreddit.
  arXiv  Detail & Related papers  (2025-01-30T01:29:46Z)
• MoralBench: Moral Evaluation of LLMs [34.43699121838648]
  This paper introduces a novel benchmark designed to measure and compare the moral reasoning capabilities of large language models (LLMs)<n>We present the first comprehensive dataset specifically curated to probe the moral dimensions of LLM outputs.<n>Our methodology involves a multi-faceted approach, combining quantitative analysis with qualitative insights from ethics scholars to ensure a thorough evaluation of model performance.
  arXiv  Detail & Related papers  (2024-06-06T18:15:01Z)
• Exploring and steering the moral compass of Large Language Models [55.2480439325792]
  Large Language Models (LLMs) have become central to advancing automation and decision-making across various sectors. This study proposes a comprehensive comparative analysis of the most advanced LLMs to assess their moral profiles.
  arXiv  Detail & Related papers  (2024-05-27T16:49:22Z)
• LLMs' Reading Comprehension Is Affected by Parametric Knowledge and Struggles with Hypothetical Statements [59.71218039095155]
  Task of reading comprehension (RC) provides a primary means to assess language models' natural language understanding (NLU) capabilities.<n>If the context aligns with the models' internal knowledge, it is hard to discern whether the models' answers stem from context comprehension or from internal information.<n>To address this issue, we suggest to use RC on imaginary data, based on fictitious facts and entities.
  arXiv  Detail & Related papers  (2024-04-09T13:08:56Z)
• SaGE: Evaluating Moral Consistency in Large Language Models [15.079905222871071]
  We show that even state-of-the-art Large Language Models are morally inconsistent in their generations. We propose an information-theoretic measure called Semantic Graph Entropy (SaGE) to measure a model's moral consistency.
  arXiv  Detail & Related papers  (2024-02-21T11:23:21Z)
• Are Large Language Models Really Robust to Word-Level Perturbations? [68.60618778027694]
  We propose a novel rational evaluation approach that leverages pre-trained reward models as diagnostic tools. Longer conversations manifest the comprehensive grasp of language models in terms of their proficiency in understanding questions. Our results demonstrate that LLMs frequently exhibit vulnerability to word-level perturbations that are commonplace in daily language usage.
  arXiv  Detail & Related papers  (2023-09-20T09:23:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.