Research

Paper

AI LLM March 11, 2026

Verbalizing LLM's Higher-order Uncertainty via Imprecise Probabilities

Authors

Anita Yang, Krikamol Muandet, Michele Caprio, Siu Lun Chau, Masaki Adachi

Abstract

Despite the growing demand for eliciting uncertainty from large language models (LLMs), empirical evidence suggests that LLM behavior is not always adequately captured by the elicitation techniques developed under the classical probabilistic uncertainty framework. This mismatch leads to systematic failure modes, particularly in settings that involve ambiguous question-answering, in-context learning, and self-reflection. To address this, we propose novel prompt-based uncertainty elicitation techniques grounded in \emph{imprecise probabilities}, a principled framework for repesenting and eliciting higher-order uncertainty. Here, first-order uncertainty captures uncertainty over possible responses to a prompt, while second-order uncertainty (uncertainty about uncertainty) quantifies indeterminacy in the underlying probability model itself. We introduce general-purpose prompting and post-processing procedures to directly elicit and quantify both orders of uncertainty, and demonstrate their effectiveness across diverse settings. Our approach enables more faithful uncertainty reporting from LLMs, improving credibility and supporting downstream decision-making.

Metadata

arXiv ID: 2603.10396
Provider: ARXIV
Primary Category: cs.AI
Published: 2026-03-11
Fetched: 2026-03-12 04:21

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.10396v1</id>\n    <title>Verbalizing LLM's Higher-order Uncertainty via Imprecise Probabilities</title>\n    <updated>2026-03-11T04:21:04Z</updated>\n    <link href='https://arxiv.org/abs/2603.10396v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.10396v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Despite the growing demand for eliciting uncertainty from large language models (LLMs), empirical evidence suggests that LLM behavior is not always adequately captured by the elicitation techniques developed under the classical probabilistic uncertainty framework. This mismatch leads to systematic failure modes, particularly in settings that involve ambiguous question-answering, in-context learning, and self-reflection. To address this, we propose novel prompt-based uncertainty elicitation techniques grounded in \\emph{imprecise probabilities}, a principled framework for repesenting and eliciting higher-order uncertainty. Here, first-order uncertainty captures uncertainty over possible responses to a prompt, while second-order uncertainty (uncertainty about uncertainty) quantifies indeterminacy in the underlying probability model itself. We introduce general-purpose prompting and post-processing procedures to directly elicit and quantify both orders of uncertainty, and demonstrate their effectiveness across diverse settings. Our approach enables more faithful uncertainty reporting from LLMs, improving credibility and supporting downstream decision-making.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-11T04:21:04Z</published>\n    <arxiv:primary_category term='cs.AI'/>\n    <author>\n      <name>Anita Yang</name>\n    </author>\n    <author>\n      <name>Krikamol Muandet</name>\n    </author>\n    <author>\n      <name>Michele Caprio</name>\n    </author>\n    <author>\n      <name>Siu Lun Chau</name>\n    </author>\n    <author>\n      <name>Masaki Adachi</name>\n    </author>\n  </entry>"
}