# Mitigating Label Bias with Interpretable Rubric Embeddings

arXiv: https://arxiv.org/abs/2605.21455 · HTML: https://mlsanity.com/paper/2605.21455/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 50 references that are in the map.
- First posted on arXiv 20 May 2026.

## Builds on (most cited first, 25 of 25)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Equality of Opportunity in Supervised Learning](https://mlsanity.com/paper/1610.02413/) (2016)
- [Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings](https://mlsanity.com/paper/1607.06520/) (2016)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Counterfactual Fairness](https://mlsanity.com/paper/1703.06856/) (2017)
- [Mitigating Unwanted Biases with Adversarial Learning](https://mlsanity.com/paper/1801.07593/) (2018)
- [Inherent Trade-Offs in the Fair Determination of Risk Scores](https://mlsanity.com/paper/1609.05807/) (2016)
- [Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection](https://mlsanity.com/paper/2004.07667/) (2020)
- [Algorithmic decision making and the cost of fairness](https://mlsanity.com/paper/1701.08230/) (2017)
- [Text and Code Embeddings by Contrastive Pre-Training](https://mlsanity.com/paper/2201.10005/) (2022)
- [Censoring Representations with an Adversary](https://mlsanity.com/paper/1511.05897/) (2015)
- [Identifying and Correcting Label Bias in Machine Learning](https://mlsanity.com/paper/1901.04966/) (2019)
- [The Measure and Mismeasure of Fairness](https://mlsanity.com/paper/1808.00023/) (2018)
- [Fair Classification with Group-Dependent Label Noise](https://mlsanity.com/paper/2011.00379/) (2020)
- [Robust and Interpretable Medical Image Classifiers via Concept Bottleneck Models](https://mlsanity.com/paper/2310.03182/) (2023)
- [Interpretable-by-Design Text Understanding with Iteratively Generated Concept Bottleneck](https://mlsanity.com/paper/2310.19660/) (2023)
- [Interpretable User Satisfaction Estimation for Conversational Systems with Large Language Models](https://mlsanity.com/paper/2403.12388/) (2024)
- [Elastic Net Regularization Paths for All Generalized Linear Models](https://mlsanity.com/paper/2103.03475/) (2021)
- [LLM-based feature generation from text for interpretable machine learning](https://mlsanity.com/paper/2409.07132/) (2024)
- [Equal Opportunity and Affirmative Action via Counterfactual Predictions](https://mlsanity.com/paper/1905.10870/) (2019)
- [Adaptive Concept Bottleneck for Foundation Models Under Distribution Shifts](https://mlsanity.com/paper/2412.14097/) (2024)
- [Risk Scores, Label Bias, and Everything but the Kitchen Sink](https://mlsanity.com/paper/2305.12638/) (2023)
- [Designing Equitable Algorithms](https://mlsanity.com/paper/2302.09157/) (2023)
- [LLMs can construct powerful representations and streamline sample-efficient supervised learning](https://mlsanity.com/paper/2603.11679/) (2026)

## Built on by (most cited first, 1 of 1)

- [From Constitutions to Control: Interpretable Rewards for Aligning Language Models](https://mlsanity.com/paper/2609.33086/) (2026)

## Abstract (arXiv)

Statistical decision algorithms are increasingly deployed in domains where ground-truth labels are hard to obtain, such as hiring, university admissions, and content moderation. In these settings, models are typically trained on historical human evaluations -- for example, using past hiring decisions as a proxy for true applicant quality. However, if past evaluations unjustly favor certain groups, models trained on these labels may inherit those biases. To address this problem, we propose basing predictions on rubric embeddings, a representation framework that replaces standard black-box embeddings with features derived from expert-defined criteria that align with the underlying construct of interest. By anchoring predictions to semantically meaningful dimensions, this approach guards against biased proxy signals. We provide both theoretical and empirical evidence that rubric embeddings mitigate label bias under plausible conditions. Empirically, we evaluate our method on a novel dataset of applications to a large master's program. We find that models trained on rubric embeddings reduce group disparities while improving measures of cohort quality. Our results suggest that basing predictions on interpretable, domain-grounded representations offers a practical approach to learning in the presence of biased labels.