# Revealed Rationality: Label-Free Evaluation and Regularization from Representation Theorems

arXiv: https://arxiv.org/abs/2608.05015 · HTML: https://mlsanity.com/paper/2608.05015/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 48 references that are in the map.
- First posted on arXiv 5 Aug 2026.

## Builds on (most cited first, 22 of 22)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Thinking Fast and Slow in Large Language Models](https://mlsanity.com/paper/2212.05206/) (2022)
- [Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs](https://mlsanity.com/paper/2502.08640/) (2025)
- [Axioms for AI Alignment from Human Feedback](https://mlsanity.com/paper/2405.14758/) (2024)
- [Beyond Preferences in AI Alignment](https://mlsanity.com/paper/2408.16984/) (2024)
- [Bayesian Teaching Enables Probabilistic Reasoning in Large Language Models](https://mlsanity.com/paper/2503.17523/) (2025)
- [Incoherent Probability Judgments in Large Language Models](https://mlsanity.com/paper/2401.16646/) (2024)
- [STEER: Assessing the Economic Rationality of Large Language Models](https://mlsanity.com/paper/2402.09552/) (2024)
- [Consistency Checks for Language Model Forecasters](https://mlsanity.com/paper/2412.18544/) (2024)
- [Scaling Open-Ended Reasoning to Predict the Future](https://mlsanity.com/paper/2512.25070/) (2025)
- [Logical Induction](https://mlsanity.com/paper/1609.03543/) (2016)
- [When Agents Say One Thing and Do Another: Validating Elicited Beliefs from LLMs](https://mlsanity.com/paper/2602.06286/) (2026)
- [Advancing Event Forecasting through Massive Training of Large Language Models: Challenges, Solutions, and Broader Impacts](https://mlsanity.com/paper/2507.19477/) (2025)
- [Can Revealed Preferences Clarify LLM Alignment and Steering?](https://mlsanity.com/paper/2605.08556/) (2026)
- [The Moral Mind(s) of Large Language Models](https://mlsanity.com/paper/2412.04476/) (2024)
- [From Drift to Coherence: Stabilizing Beliefs in LLMs](https://mlsanity.com/paper/2606.17832/) (2026)
- [New developments in revealed preference theory: decisions under risk, uncertainty, and intertemporal choice](https://mlsanity.com/paper/1908.07561/) (2019)
- [Incoherent Values? Probing LLM Preferences Through Parametric Variation](https://mlsanity.com/paper/2606.21102/) (2026)
- [Economic Rationality under Specialization: Evidence of Decision Bias in AI Agents](https://mlsanity.com/paper/2501.18190/) (2025)
- [Sparks of Rationality: Do Reasoning LLMs Align with Human Judgment and Choice?](https://mlsanity.com/paper/2601.22329/) (2026)

## Built on by (most cited first, 3 of 3)

- [Dutch Books for Language Models](https://mlsanity.com/paper/2609.02797/) (2026)
- [Probability Contracts: Accuracy, Coherence, and Decisions Across LLM Interfaces](https://mlsanity.com/paper/2609.37470/) (2026)
- [Understanding as No-Arbitrage: Bounded Dutch Books as a Definition and Training Objective for Language Models](https://mlsanity.com/paper/2609.39341/) (2026)

## Abstract (arXiv)

Representation theorems in decision theory establish that behavior satisfies certain axioms if and only if it can be rationalized by a well-defined objective. I argue that this ``if and only if'' structure provides a potentially useful foundation for label-free evaluation and regularization of LLMs and other AI systems. Axiom compliance can be checked from the model's own responses to synthetic choice problems, with no external labels or human feedback, and the penalties are readily computable. Because the axioms are necessary and sufficient, the resulting checks exhaust the implications of the relevant rationality standard for the elicited data: a model that passes cannot be rejected on rationality grounds by any further test of the same data. I discuss three instantiations: probabilistic coherence via a theorem of de Finetti, preference rationality via Afriat's theorem, and subjective expected utility via a theorem of Echenique and Saito (2015), each yielding a continuous penalty that is zero whenever behavior can be rationalized. Since coherence does not restrict which objective rationalizes behavior, these penalties complement rather than replace other evaluation and training signals.