# Self-Consistency of Large Language Models under Ambiguity

arXiv: https://arxiv.org/abs/2310.13439 · HTML: https://mlsanity.com/paper/2310.13439/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 14 of its 14 references that are in the map.
- First posted on arXiv 20 Oct 2023.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 14 of 14)

- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting](https://mlsanity.com/paper/2305.04388/) (2023)
- [Whose Opinions Do Language Models Reflect?](https://mlsanity.com/paper/2303.17548/) (2023)
- [Discovering Latent Knowledge in Language Models Without Supervision](https://mlsanity.com/paper/2212.03827/) (2022)
- [Teaching Models to Express Their Uncertainty in Words](https://mlsanity.com/paper/2205.14334/) (2022)
- [Measuring and Improving Consistency in Pretrained Language Models](https://mlsanity.com/paper/2102.01017/) (2021)
- [We're Afraid Language Models Aren't Modeling Ambiguity](https://mlsanity.com/paper/2304.14399/) (2023)
- [Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations](https://mlsanity.com/paper/2307.08678/) (2023)
- [Still No Lie Detector for Language Models: Probing Empirical and Conceptual Roadblocks](https://mlsanity.com/paper/2307.00175/) (2023)
- [Evaluating Superhuman Models with Consistency Checks](https://mlsanity.com/paper/2306.09983/) (2023)
- [Measuring Reliability of Large Language Models through Semantic Consistency](https://mlsanity.com/paper/2211.05853/) (2022)
- [Task Ambiguity in Humans and Language Models](https://mlsanity.com/paper/2212.10711/) (2022)

## Built on by (most cited first, 12 of 12)

- [Knowledge Conflicts for LLMs: A Survey](https://mlsanity.com/paper/2403.08319/) (2024)
- [The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs](https://mlsanity.com/paper/2501.10970/) (2025)
- [Unraveling Cross-Modality Knowledge Conflicts in Large Vision-Language Models](https://mlsanity.com/paper/2410.03659/) (2024)
- [Semantic Energy: Detecting LLM Hallucination Beyond Entropy](https://mlsanity.com/paper/2508.14496/) (2025)
- [Capturing Polysemanticity with PRISM: A Multi-Concept Feature Description Framework](https://mlsanity.com/paper/2506.15538/) (2025)
- [Text Speaks Louder than Vision: ASCII Art Reveals Textual Biases in Vision-Language Models](https://mlsanity.com/paper/2504.01589/) (2025)
- [How does Misinformation Affect Large Language Model Behaviors and Preferences?](https://mlsanity.com/paper/2505.21608/) (2025)
- [Is Cognition Consistent with Perception? Assessing and Mitigating Multimodal Knowledge Conflicts in Document Understanding](https://mlsanity.com/paper/2411.07722/) (2024)
- [Interpreting Language Models Through Concept Descriptions: A Survey](https://mlsanity.com/paper/2510.01048/) (2025)
- [Generative AI for Managerial Decision-Making under Ambiguity and Sycophancy](https://mlsanity.com/paper/2603.03970/) (2026)
- [Learning by Teaching: Engaging Students as Instructors of Large Language Models in Computer Science Education](https://mlsanity.com/paper/2508.05979/) (2025)
- [Hallucinations in LLMs: A Lifecycle-Based Survey of Causes, Detection, Mitigation, and Prevention](https://mlsanity.com/paper/2608.26168/) (2026)

## Abstract (arXiv)

Large language models (LLMs) that do not give consistent answers across contexts are problematic when used for tasks with expectations of consistency, e.g., question-answering, explanations, etc. Our work presents an evaluation benchmark for self-consistency in cases of under-specification where two or more answers can be correct. We conduct a series of behavioral experiments on the OpenAI model suite using an ambiguous integer sequence completion task. We find that average consistency ranges from 67% to 82%, far higher than would be predicted if a model's consistency was random, and increases as model capability improves. Furthermore, we show that models tend to maintain self-consistency across a series of robustness checks, including prompting speaker changes and sequence length changes. These results suggest that self-consistency arises as an emergent capability without specifically training for it. Despite this, we find that models are uncalibrated when judging their own consistency, with models displaying both over- and under-confidence. We also propose a nonparametric test for determining from token output distribution whether a model assigns non-trivial probability to alternative answers. Using this test, we find that despite increases in self-consistency, models usually place significant weight on alternative, inconsistent answers. This distribution of probability mass provides evidence that even highly self-consistent models internally compute multiple possible responses.