Failure of contextual invariance in gender inference with large language models

• URL: http://arxiv.org/abs/2603.23485v1
• Date: Tue, 24 Mar 2026 17:52:22 GMT
• Title: Failure of contextual invariance in gender inference with large language models
• Abstract summary: We introduce minimal, theoretically uninformative discourse context and find that this induces large, systematic shifts in model outputs.<n> Correlations with cultural gender stereotypes, present in decontextualized settings, weaken or disappear once context is introduced.<n>A Contextuality-by-Default analysis reveals that, in 19--52% of cases across models, this dependence persists.
• Score: 0.143494686131174
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Standard evaluation practices assume that large language model (LLM) outputs are stable under contextually equivalent formulations of a task. Here, we test this assumption in the setting of gender inference. Using a controlled pronoun selection task, we introduce minimal, theoretically uninformative discourse context and find that this induces large, systematic shifts in model outputs. Correlations with cultural gender stereotypes, present in decontextualized settings, weaken or disappear once context is introduced, while theoretically irrelevant features, such as the gender of a pronoun for an unrelated referent, become the most informative predictors of model behaviour. A Contextuality-by-Default analysis reveals that, in 19--52\% of cases across models, this dependence persists after accounting for all marginal effects of context on individual outputs and cannot be attributed to simple pronoun repetition. These findings show that LLM outputs violate contextual invariance even under near-identical syntactic formulations, with implications for bias benchmarking and deployment in high-stakes settings.

Related papers

• The Illusion of Robustness: Aggregate Accuracy Hides Prediction Flips under Task-Irrelevant Context [76.36635039699112]
  Large language models (LLMs) are increasingly deployed in context-rich settings where task inputs are often accompanied by long, partially irrelevant context.<n>We find that state-of-the-art models often appear robust to task-irrelevant context at the aggregate level.<n>This aggregate stability, however, masks significant per-example instability.
  arXiv  Detail & Related papers  (2026-07-14T17:01:12Z)
• Alignment Reduces Expressed but Not Encoded Gender Bias: A Unified Framework and Study [3.679036235271287]
  We propose a unified framework to jointly analyze intrinsic and extrinsic gender bias in Large Language Models.<n>We find a consistent association between latent gender information and expressed bias when measured under the unified protocol.<n>Our results suggest that while the latter indeed reduces expressed bias, measurable gender-related associations are still present in internal representations.
  arXiv  Detail & Related papers  (2026-03-25T09:35:18Z)
• Context-Aware Counterfactual Data Augmentation for Gender Bias Mitigation in Language Models [14.937360204897415]
  A challenge in mitigating social bias in fine-tuned language models (LMs) is the potential reduction in language modeling capability.<n>We propose a context-augmented data augmentation method, Context-CDA, which uses large LMs to enhance the diversity and contextual relevance of the debiasing corpus.<n>We then employ uncertainty-based filtering to exclude generated counterfactuals considered low-quality by the target smaller LMs.
  arXiv  Detail & Related papers  (2026-02-10T09:45:21Z)
• What Triggers my Model? Contrastive Explanations Inform Gender Choices by Translation Models [0.0]
  Interpretability can be implemented as a means to understand decisions taken by (black box) models, such as machine translation (MT) or large language models (LLMs)<n>Yet, research in this area has been limited in relation to a manifested problem in these models: gender bias.<n>This study examines which context, in the form of input tokens in the source sentence, influences (or triggers) the translation model choice of a certain gender inflection in the target language.
  arXiv  Detail & Related papers  (2025-12-09T10:14:10Z)
• Uncertainty Quantification for Evaluating Machine Translation Bias [8.210153370613941]
  We use uncertainty to measure gender bias in machine translation systems.<n>We find that high translation accuracy does not correlate with exhibiting uncertainty appropriately, and that debiasing affects the two cases differently.
  arXiv  Detail & Related papers  (2025-07-24T12:10:21Z)
• Assumed Identities: Quantifying Gender Bias in Machine Translation of Gender-Ambiguous Occupational Terms [12.568906647547815]
  We introduce GRAPE, a probability-based metric designed to evaluate gender bias.<n>We present GAMBIT, a benchmarking dataset in English with gender-ambiguous occupational terms.<n>Using GRAPE, we evaluate several MT systems and examine whether their gendered translations in Greek and French align with or diverge from societal stereotypes.
  arXiv  Detail & Related papers  (2025-03-06T12:16:14Z)
• Beyond Binary Gender: Evaluating Gender-Inclusive Machine Translation with Ambiguous Attitude Words [85.48043537327258]
  Existing machine translation gender bias evaluations are primarily focused on male and female genders. This study presents a benchmark AmbGIMT (Gender-Inclusive Machine Translation with Ambiguous attitude words) We propose a novel process to evaluate gender bias based on the Emotional Attitude Score (EAS), which is used to quantify ambiguous attitude words.
  arXiv  Detail & Related papers  (2024-07-23T08:13:51Z)
• All Should Be Equal in the Eyes of Language Models: Counterfactually Aware Fair Text Generation [16.016546693767403]
  We propose a framework that dynamically compares the model understanding of diverse demographics to generate more equitable sentences. CAFIE produces fairer text and strikes the best balance between fairness and language modeling capability.
  arXiv  Detail & Related papers  (2023-11-09T15:39:40Z)
• A Tale of Pronouns: Interpretability Informs Gender Bias Mitigation for Fairer Instruction-Tuned Machine Translation [35.44115368160656]
  We investigate whether and to what extent machine translation models exhibit gender bias. We find that IFT models default to male-inflected translations, even disregarding female occupational stereotypes. We propose an easy-to-implement and effective bias mitigation solution.
  arXiv  Detail & Related papers  (2023-10-18T17:36:55Z)
• How to Handle Different Types of Out-of-Distribution Scenarios in Computational Argumentation? A Comprehensive and Fine-Grained Field Study [59.13867562744973]
  This work systematically assesses LMs' capabilities for out-of-distribution (OOD) scenarios. We find that the efficacy of such learning paradigms varies with the type of OOD. Specifically, while ICL excels for domain shifts, prompt-based fine-tuning surpasses for topic shifts.
  arXiv  Detail & Related papers  (2023-09-15T11:15:47Z)
• Counter-GAP: Counterfactual Bias Evaluation through Gendered Ambiguous Pronouns [53.62845317039185]
  Bias-measuring datasets play a critical role in detecting biased behavior of language models. We propose a novel method to collect diverse, natural, and minimally distant text pairs via counterfactual generation. We show that four pre-trained language models are significantly more inconsistent across different gender groups than within each group.
  arXiv  Detail & Related papers  (2023-02-11T12:11:03Z)
• Causally Testing Gender Bias in LLMs: A Case Study on Occupational Bias [33.99768156365231]
  We introduce a causal formulation for bias measurement in generative language models.<n>We propose a benchmark called OccuGender, with a bias-measuring procedure to investigate occupational gender bias.<n>The results show that these models exhibit substantial occupational gender bias.
  arXiv  Detail & Related papers  (2022-12-20T22:41:24Z)
• Naturalistic Causal Probing for Morpho-Syntax [76.83735391276547]
  We suggest a naturalistic strategy for input-level intervention on real world data in Spanish. Using our approach, we isolate morpho-syntactic features from counfounders in sentences. We apply this methodology to analyze causal effects of gender and number on contextualized representations extracted from pre-trained models.
  arXiv  Detail & Related papers  (2022-05-14T11:47:58Z)
• Balancing out Bias: Achieving Fairness Through Training Reweighting [58.201275105195485]
  Bias in natural language processing arises from models learning characteristics of the author such as gender and race. Existing methods for mitigating and measuring bias do not directly account for correlations between author demographics and linguistic variables. This paper introduces a very simple but highly effective method for countering bias using instance reweighting.
  arXiv  Detail & Related papers  (2021-09-16T23:40:28Z)
• The Gap on GAP: Tackling the Problem of Differing Data Distributions in Bias-Measuring Datasets [58.53269361115974]
  Diagnostic datasets that can detect biased models are an important prerequisite for bias reduction within natural language processing. undesired patterns in the collected data can make such tests incorrect. We introduce a theoretically grounded method for weighting test samples to cope with such patterns in the test data.
  arXiv  Detail & Related papers  (2020-11-03T16:50:13Z)

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.