Inverse Constitutional AI: Compressing Preferences into Principles

Arduin Findeis, Timo Kaufmann, Eyke Hüllermeier, Samuel Albanie, Robert D. Mullins

Keywords: human feedback, evaluation, interpretability, preference learning, AI annotators

TL;DR: We compress pairwise text preferences into a set of interpretable principles (a constitution) that may help identify undesirable biases, improve AI annotators and assist in personalizing LLMs.

Abstract: Feedback data is widely used for fine-tuning and evaluating state-of-the-art AI models. Pairwise text preferences, where human or AI annotators select the “better” of two options, are particularly common. Such preferences are used to train (reward) models or to rank models with aggregate statistics. For many applications it is desirable to understand annotator preferences in addition to modelling them  – not least because extensive prior work has shown various unintended biases in preference datasets. Yet, preference datasets remain challenging to interpret. Neither black-box reward models nor statistics can answer why one text is preferred over another. Manual interpretation of the numerous (long) response pairs is usually equally infeasible. In this paper, we introduce the Inverse Constitutional AI (ICAI) problem, formulating the interpretation of pairwise text preference data as a compression task. In constitutional AI, a set of principles (a constitution) is used to provide feedback and fine-tune AI models. ICAI inverts this process: given a feedback dataset, we aim to extract a constitution that best enables a large language model (LLM) to reconstruct the original annotations. We propose a corresponding ICAI algorithm and validate its generated constitutions quantitatively based on annotation reconstruction accuracy on several datasets: (a) synthetic feedback data with known principles; (b) AlpacaEval cross-annotated human feedback data; (c) crowdsourced Chatbot Arena data; and (d) PRISM data from diverse demographic groups. As an example application, we further demonstrate the detection of biases in human feedback data. As a short and interpretable representation of the original dataset, generated constitutions have many potential use cases: they may help identify undesirable annotator biases, better understand model performance, scale feedback to unseen data, or assist with adapting AI models to individual user or group preferences. We release the source code for our algorithm and experiments at https://github.com/rdnfn/icai.

Primary Area: alignment, fairness, safety, privacy, and societal considerations

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 6631