OMIT the Action: Measuring Framing-Invariant Omission Bias under Philosophical Disagreement

Key points

• As LLMs increasingly assist in moral reasoning, omission bias, the tendency to prefer inaction even when equivalent framings reverse substantive outcomes, poses a significant risk of skewed decision-making.1
• Evaluating eight LLMs, we find that omission bias is pervasive but inversely correlates with model size within families.1
• We further evaluate four inference-time interventions and find that interventions encouraging models to consider moral principles before committing to a yes/no answer reduce omission bias and increase frame-consistent responses, although lower omission bias rates can also coincide with shifts toward action-biased responses.1
• Ultimately, this work contributes not only the OMIT benchmark, but also a methodology for using diverse philosophical disagreement signals to evaluate framing-sensitive inaction preferences and the distributional effects of mitigation attempts in LLMs under complex moral conflicts.1

Sources (1)

• [1]OMIT the Action: Measuring Framing-Invariant Omission Bias under Philosophical Disagreement
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 06:54 AM

  To address this gap, we introduce OMIT, a benchmark consisting of 218 paired-frame scenarios across 10 conflict types, constructed by leveraging disagreement patterns from an LLM-based, five-perspective philosophical persona panel (utilitarianism, deontology, virtue ethics, care ethics, and contractualism).

  As LLMs increasingly assist in moral reasoning, omission bias, the tendency to prefer inaction even when equivalent framings reverse substantive outcomes, poses a significant risk of skewed decision-making.

Extractive summary: sentences quoted from the sources.