On Defining Erasure Harms for NLP

• URL: http://arxiv.org/abs/2606.15815v1
• Date: Sun, 14 Jun 2026 13:34:10 GMT
• Title: On Defining Erasure Harms for NLP
• Abstract summary: Deployment of NLP systems has raised concerns about harms they might produce.<n>Recent literature has begun to conceptualize and measure one such harm, the harm of erasure.<n>We propose a structured definition of erasure that clarifies what components are necessary for establishing whether erasure has occurred.
• Score: 69.37262811969147
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The deployment of NLP systems has raised concerns about harms they might produce, including representational harms. Recent literature has begun to conceptualize and measure one such harm, the harm of erasure. Nevertheless, the field lacks a clear and cohesive conceptual foundation for identifying and measuring erasure. Existing conceptualizations of erasure are often broad -- making it difficult to identify what is needed to establish and measure erasure -- or else specific to particular settings -- facilitating measurement for those settings but potentially challenging to adapt to other settings. To address this gap, we develop and propose a structured definition of erasure that clarifies what components are necessary for establishing whether erasure has occurred, which practitioners need to explicitly articulate and operationalize in order to measure erasure.

Related papers

• Are We Evaluating the Edit Locality of LLM Model Editing Properly? [68.441768731381]
  We find that existing specificity evaluation protocols are inadequate for this purpose.<n>Existing specificity metrics are weakly correlated with the strength of specificity regularizers.<n>We also find that current metrics lack sufficient sensitivity, rendering them ineffective at distinguishing the specificity performance of different methods.
  arXiv  Detail & Related papers  (2026-01-24T07:07:21Z)
• SAGE: Exploring the Boundaries of Unsafe Concept Domain with Semantic-Augment Erasing [65.82241040239452]
  Concept erasing finetunes weights to unlearn undesirable concepts.<n>Existing methods treat unsafe concept as a fixed word and repeatedly erase it.<n>We introduce semantic-augment erasing which transforms concept word erasure into concept domain erasure.
  arXiv  Detail & Related papers  (2025-06-11T03:21:24Z)
• Does Representation Intervention Really Identify Desired Concepts and Elicit Alignment? [73.80382983108997]
  Representation intervention aims to locate and modify the representations that encode the underlying concepts in Large Language Models.<n>If the interventions are faithful, the intervened LLMs should erase the harmful concepts and be robust to both in-distribution adversarial prompts and the out-of-distribution jailbreaks.<n>We propose Concept Concentration (COCA), which simplifies the decision boundary between harmful and benign representations.
  arXiv  Detail & Related papers  (2025-05-24T12:23:52Z)
• SConU: Selective Conformal Uncertainty in Large Language Models [59.25881667640868]
  We propose a novel approach termed Selective Conformal Uncertainty (SConU)<n>We develop two conformal p-values that are instrumental in determining whether a given sample deviates from the uncertainty distribution of the calibration set at a specific manageable risk level.<n>Our approach not only facilitates rigorous management of miscoverage rates across both single-domain and interdisciplinary contexts, but also enhances the efficiency of predictions.
  arXiv  Detail & Related papers  (2025-04-19T03:01:45Z)
• Rethinking the Uncertainty: A Critical Review and Analysis in the Era of Large Language Models [42.563558441750224]
  Large Language Models (LLMs) have become fundamental to a broad spectrum of artificial intelligence applications. Current methods often struggle to accurately identify, measure, and address the true uncertainty. This paper introduces a comprehensive framework specifically designed to identify and understand the types and sources of uncertainty.
  arXiv  Detail & Related papers  (2024-10-26T15:07:15Z)
• Label-wise Aleatoric and Epistemic Uncertainty Quantification [15.642370299038488]
  We present a novel approach to uncertainty quantification in classification tasks based on label-wise decomposition of uncertainty measures. We show that our proposed measures adhere to a number of desirable properties.
  arXiv  Detail & Related papers  (2024-06-04T14:33:23Z)
• Segmentation Re-thinking Uncertainty Estimation Metrics for Semantic Segmentation [12.532289778772185]
  semantic segmentation is a fundamental application within machine learning. The metric known as PAvPU (Patch Accuracy versus Patch Uncertainty) has been developed as a specialized tool for evaluating entropy-based uncertainty in image segmentation tasks. Our investigation identifies three core deficiencies within the PAvPU framework and proposes robust solutions.
  arXiv  Detail & Related papers  (2024-03-28T20:34:02Z)
• Beyond Behaviorist Representational Harms: A Plan for Measurement and Mitigation [1.7355698649527407]
  This study focuses on an examination of current definitions of representational harms to discern what is included and what is not. Our work highlights the unique vulnerabilities of large language models to perpetrating representational harms. The overarching aim of this research is to establish a framework for broadening the definition of representational harms.
  arXiv  Detail & Related papers  (2024-01-25T00:54:10Z)
• Tailored Uncertainty Estimation for Deep Learning Systems [10.288326973530614]
  We propose a framework that guides the selection of a suitable uncertainty estimation method. Our framework provides strategies to validate this choice and to uncover structural weaknesses. It anticipates prospective machine learning regulations that require evidences for the technical appropriateness of machine learning systems.
  arXiv  Detail & Related papers  (2022-04-29T09:23:07Z)
• Localization Uncertainty Estimation for Anchor-Free Object Detection [48.931731695431374]
  There are several limitations of the existing uncertainty estimation methods for anchor-based object detection. We propose a new localization uncertainty estimation method called UAD for anchor-free object detection. Our method captures the uncertainty in four directions of box offsets that are homogeneous, so that it can tell which direction is uncertain.
  arXiv  Detail & Related papers  (2020-06-28T13:49:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.