Confident, Calibrated, or Complicit: Probing the Trade-offs between Safety Alignment and Ideological Bias in Language Models in Detecting Hate Speech

• URL: http://arxiv.org/abs/2509.00673v1
• Date: Sun, 31 Aug 2025 03:00:55 GMT
• Title: Confident, Calibrated, or Complicit: Probing the Trade-offs between Safety Alignment and Ideological Bias in Language Models in Detecting Hate Speech
• Authors: Sanjeeevan Selvaganapathy, Mehwish Nasim,
• Abstract summary: We investigate the efficacy of Large Language Models (LLMs) in detecting implicit and explicit hate speech.<n>We find that censored models significantly outperform their uncensored counterparts in both accuracy and robustness.
• Score: 0.916708284510944
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: We investigate the efficacy of Large Language Models (LLMs) in detecting implicit and explicit hate speech, examining whether models with minimal safety alignment (uncensored) might provide more objective classification capabilities compared to their heavily-aligned (censored) counterparts. While uncensored models theoretically offer a less constrained perspective free from moral guardrails that could bias classification decisions, our results reveal a surprising trade-off: censored models significantly outperform their uncensored counterparts in both accuracy and robustness, achieving 78.7% versus 64.1% strict accuracy. However, this enhanced performance comes with its own limitation -- the safety alignment acts as a strong ideological anchor, making censored models resistant to persona-based influence, while uncensored models prove highly malleable to ideological framing. Furthermore, we identify critical failures across all models in understanding nuanced language such as irony. We also find alarming fairness disparities in performance across different targeted groups and systemic overconfidence that renders self-reported certainty unreliable. These findings challenge the notion of LLMs as objective arbiters and highlight the need for more sophisticated auditing frameworks that account for fairness, calibration, and ideological consistency.

Related papers

• Can Large Language Models Express Uncertainty Like Human? [71.27418419522884]
  We release the first diverse, large-scale dataset of hedging expressions with human-annotated confidence scores.<n>We conduct the first systematic study of linguistic confidence across modern large language models.
  arXiv  Detail & Related papers  (2025-09-29T02:34:30Z)
• Investigating Intersectional Bias in Large Language Models using Confidence Disparities in Coreference Resolution [5.061421107401101]
  Large language models (LLMs) have achieved impressive performance, leading to their widespread adoption as decision-support tools in resource-constrained contexts like hiring and admissions.<n>There is, however, scientific consensus that AI systems can reflect and exacerbate societal biases, raising concerns about identity-based harm when used in critical social contexts.<n>In this work, we extend single-axis fairness evaluations to examine intersectional bias, recognizing that when multiple axes of discrimination intersect, they create distinct patterns of disadvantage.
  arXiv  Detail & Related papers  (2025-08-09T22:24:40Z)
• MetaFaith: Faithful Natural Language Uncertainty Expression in LLMs [66.14178164421794]
  We introduce MetaFaith, a novel prompt-based calibration approach inspired by human metacognition.<n>We show that MetaFaith robustly improves faithful calibration across diverse models and task domains, enabling up to 61% improvement in faithfulness.
  arXiv  Detail & Related papers  (2025-05-30T17:54:08Z)
• Confidential Guardian: Cryptographically Prohibiting the Abuse of Model Abstention [65.47632669243657]
  A dishonest institution can exploit mechanisms to discriminate or unjustly deny services under the guise of uncertainty.<n>We demonstrate the practicality of this threat by introducing an uncertainty-inducing attack called Mirage.<n>We propose Confidential Guardian, a framework that analyzes calibration metrics on a reference dataset to detect artificially suppressed confidence.
  arXiv  Detail & Related papers  (2025-05-29T19:47:50Z)
• Language Models That Walk the Talk: A Framework for Formal Fairness Certificates [6.5301153208275675]
  This work presents a holistic verification framework to certify the robustness of transformer-based language models.<n>We focus on ensuring gender fairness and consistent outputs across different gender-related terms.<n>We extend this methodology to toxicity detection, offering formal guarantees that adversarially manipulated toxic inputs are consistently detected and appropriately censored.
  arXiv  Detail & Related papers  (2025-05-19T06:46:17Z)
• Benchmarking Adversarial Robustness to Bias Elicitation in Large Language Models: Scalable Automated Assessment with LLM-as-a-Judge [1.1666234644810893]
  Small models outperform larger ones in safety, suggesting that training and architecture may matter more than scale.<n>No model is fully robust to adversarial elicitation, with jailbreak attacks using low-resource languages or refusal suppression proving effective.
  arXiv  Detail & Related papers  (2025-04-10T16:00:59Z)
• Turning Logic Against Itself : Probing Model Defenses Through Contrastive Questions [51.51850981481236]
  We introduce POATE, a novel jailbreak technique that harnesses contrastive reasoning to provoke unethical responses.<n>PoATE crafts semantically opposing intents and integrates them with adversarial templates, steering models toward harmful outputs with remarkable subtlety.<n>To counter this, we propose Intent-Aware CoT and Reverse Thinking CoT, which decompose queries to detect malicious intent and reason in reverse to evaluate and reject harmful responses.
  arXiv  Detail & Related papers  (2025-01-03T15:40:03Z)
• On the Fairness, Diversity and Reliability of Text-to-Image Generative Models [68.62012304574012]
  multimodal generative models have sparked critical discussions on their reliability, fairness and potential for misuse.<n>We propose an evaluation framework to assess model reliability by analyzing responses to global and local perturbations in the embedding space.<n>Our method lays the groundwork for detecting unreliable, bias-injected models and tracing the provenance of embedded biases.
  arXiv  Detail & Related papers  (2024-11-21T09:46:55Z)
• Covert Bias: The Severity of Social Views' Unalignment in Language Models Towards Implicit and Explicit Opinion [0.40964539027092917]
  We evaluate the severity of bias toward a view by using a biased model in edge cases of excessive bias scenarios. Our findings reveal a discrepancy in LLM performance in identifying implicit and explicit opinions, with a general tendency of bias toward explicit opinions of opposing stances. The direct, incautious responses of the unaligned models suggest a need for further refinement of decisiveness.
  arXiv  Detail & Related papers  (2024-08-15T15:23:00Z)
• Don't Go To Extremes: Revealing the Excessive Sensitivity and Calibration Limitations of LLMs in Implicit Hate Speech Detection [29.138463029748547]
  This paper explores the capability of Large Language Models to detect implicit hate speech and express confidence in their responses. Our findings highlight that LLMs exhibit two extremes: (1) LLMs display excessive sensitivity towards groups or topics that may cause fairness issues, resulting in misclassifying benign statements as hate speech.
  arXiv  Detail & Related papers  (2024-02-18T00:04:40Z)
• Characterizing the adversarial vulnerability of speech self-supervised learning [95.03389072594243]
  We make the first attempt to investigate the adversarial vulnerability of such paradigm under the attacks from both zero-knowledge adversaries and limited-knowledge adversaries. The experimental results illustrate that the paradigm proposed by SUPERB is seriously vulnerable to limited-knowledge adversaries.
  arXiv  Detail & Related papers  (2021-11-08T08:44:04Z)
• Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models [86.02610674750345]
  Adversarial GLUE (AdvGLUE) is a new multi-task benchmark to explore and evaluate the vulnerabilities of modern large-scale language models under various types of adversarial attacks. We apply 14 adversarial attack methods to GLUE tasks to construct AdvGLUE, which is further validated by humans for reliable annotations. All the language models and robust training methods we tested perform poorly on AdvGLUE, with scores lagging far behind the benign accuracy.
  arXiv  Detail & Related papers  (2021-11-04T12:59:55Z)

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.