mlsanityData to 14 Oct 2026

EMBRACE: Shaping Inclusive Opinion Representation by Aligning Implicit Conversations with Social Norms

arXiv 2507.20264 ↗First posted 27 Jul 2025cs.CLcited by 1

Not on the map. Why?

Abstract (arXiv)

Shaping inclusive representations that embrace diversity and ensure fair participation and reflections of values is at the core of many conversation-based models. However, many existing methods rely on surface inclusion using mention of user demographics or behavioral attributes of social groups. Such methods overlook the nuanced, implicit expression of opinion embedded in conversations. Furthermore, the over-reliance on overt cues can exacerbate misalignment and reinforce harmful or stereotypical representations in model outputs. Thus, we took a step back and recognized that equitable inclusion needs to account for the implicit expression of opinion and use the stance of responses to validate the normative alignment. This study aims to evaluate how opinions are represented in NLP or computational models by introducing an alignment evaluation framework that foregrounds implicit, often overlooked conversations and evaluates the normative social views and discourse. Our approach models the stance of responses as a proxy for the underlying opinion, enabling a considerate and reflective representation of diverse social viewpoints. We evaluate the framework using both (i) positive-unlabeled (PU) online learning with base classifiers, and (ii) instruction-tuned language models to assess post-training alignment. Through this, we provide a principled and structured lens on how implicit opinions are (mis)represented and offer a pathway toward more inclusive model behavior.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 50 references that are in the map. First posted on arXiv 27 Jul 2025.

Builds on

24 of 50 references.

References · 24

  1. 2024WildChat: 1M ChatGPT Interaction Logs in the Wildcited by 629
  2. 2019Social Bias Frames: Reasoning about Social and Power Implications of Languagecited by 380
  3. 2023ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversationcited by 231
  4. 2021Latent Hatred: A Benchmark for Understanding Implicit Hate Speechcited by 217
  5. 2024Modular Pluralism: Pluralistic Alignment via Multi-LLM Collaborationcited by 113
  6. 2025Measuring Sycophancy of Language Models in Multi-turn Dialoguescited by 104
  7. 2020Stance Detection on Social Media: State of the Art and Trendscited by 95
  8. 2024Unintended Impacts of LLM Alignment on Global Representationcited by 66
  9. 2024Towards Implicit Bias Detection and Mitigation in Multi-Agent LLM Interactionscited by 51
  10. 2022Human-Machine Collaboration Approaches to Build a Dialogue Dataset for Hate Speech Counteringcited by 40
  11. 2024From Distributional to Overton Pluralism: Investigating Large Language Model Alignmentcited by 33
  12. 2025Unmasking Implicit Bias: Evaluating Persona-Prompted LLM Responses in Power-Disparate Social Scenarioscited by 26
12 more
  1. 2017The Role of Conversation Context for Sarcasm Detection in Online Interactionscited by 16
  2. 2025Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Racecited by 15
  3. 2023Chain-of-Thought Embeddings for Stance Detection on Social Mediacited by 13
  4. 2025Labeling Messages as AI-Generated Does Not Reduce Their Persuasive Effectscited by 13
  5. 2024Subtle Biases Need Subtler Measures: Dual Metrics for Evaluating Representative and Affinity Bias in Large Language Modelscited by 10
  6. 2025Aligning Large Language Models with Implicit Preferences from User-Generated Contentcited by 9
  7. 2024Evaluating Implicit Bias in Large Language Models by Attacking From a Psychometric Perspectivecited by 6
  8. 2024Superlatives in Context: Modeling the Implicit Semantics of Superlativescited by 5
  9. 2025What's the Difference? Supporting Users in Identifying the Effects of Prompt and Model Changes Through Token Patternscited by 4
  10. 2024Covert Bias: The Severity of Social Views' Unalignment in Language Models Towards Implicit and Explicit Opinioncited by 3
  11. 2023Guiding Computational Stance Detection with Expanded Stance Triangle Frameworkcited by 2
  12. 2024USDC: A Dataset of underlineUser underlineStance and underlineDogmatism in Long underlineConversationscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Same Lesson, Different Story: Cross-Lingual Reconstruction of Cultural Narratives in Large Language Modelscited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2025MINDS: A Cross-cultural Dialogue Corpus for Social Norm Classification and Adherence Detection
  2. 2025Surfacing Subtle Stereotypes: A Multilingual, Debate-Oriented Evaluation of Modern LLMs
  3. 2024The Root Shapes the Fruit: On the Persistence of Gender-Exclusive Harms in Aligned Language Models
  4. 2024COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities
  5. 2025Examining Alignment of Large Language Models through Representative Heuristics: The Case of Political Stereotypes
  6. 2026Edu-MMBias: A Three-Tier Multimodal Benchmark for Auditing Social Bias in Vision-Language Models under Educational Contexts
  7. 2024The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models
  8. 2025Explicit vs. Implicit: Investigating Social Bias in Large Language Models through Self-Reflection