Generating spoken dialogue is inherently more complex than monologue text-to-speech (TTS), as it demands both realistic turn-taking and the maintenance of distinct speaker timbres. While existing autoregressive (AR) models have made progress, they often suffer from high inference latency and stability issues. To overcome these limitations, we propose ZipVoice-Dialog, a non-autoregressive (NAR) zero-shot spoken dialogue generation model based on flow-matching. Observing that applying vanilla flow-matching to dialogue generation leads to poor speech intelligibility and turn-taking precision, we introduce two simple yet effective methods to adapt flow-matching architectures for dialogue generation: (1) a curriculum learning strategy to ensure robust speech-text alignment, and (2) speaker-turn embeddings to govern precise speaker turn-taking. Additionally, we introduce dedicated strategies to support stereo dialogue generation.Recognizing the lack of training datasets in this field, we curate and release OpenDialog, the first large-scale (6.8k hours) open-source spoken dialogue dataset derived from in-the-wild speech data. Moreover, for fair and rigorous evaluations, we established a benchmark to comprehensively evaluate dialogue generation models. Experiments demonstrate the effectiveness of the proposed methods and dataset, showing that ZipVoice-Dialog achieves superior performance in inference speed, intelligibility, speaker turn-taking accuracy, and speaker similarity. Our code, model checkpoints, and the OpenDialog dataset are publicly available.
论文检索
输入标题、作者或关键词，从 700 篇学术成果中精准定位
Analysing Differences in Persuasive Language in LLM-Generated Text: Uncovering Stereotypical Gender Patterns
PDF ↗Large language models (LLMs) are increasingly used for everyday communication tasks, including drafting interpersonal messages intended to influence and persuade. Prior work has shown that LLMs can successfully persuade humans and amplify persuasive language. It is therefore essential to understand how user instructions affect the generation of persuasive language, and to understand whether the generated persuasive language differs, for example, when targeting different groups. In this work, we propose a framework for evaluating how persuasive language generation is affected by recipient gender, sender intent, or output language. We evaluate 13 LLMs and 16 languages using pairwise prompt instructions. We evaluate model responses on 19 categories of persuasive language using an LLM-as-judge setup grounded in social psychology and communication science. Our results reveal significant gender differences in the persuasive language generated across all models. These patterns reflect biases consistent with gender-stereotypical linguistic tendencies documented in social psychology and sociolinguistics.
As high-quality data becomes harder to obtain, reward models are increasingly important. Beyond the costly RLHF stage, they are now used at inference time to guide LLM generation and in data selection for post-training. These methods bring efficiency and performance gains, but current reward models often fail to prevent untrustworthy behaviors such as privacy leaks and stereotypes. Re-training reward models to address these issues is expensive, since it requires large-scale human preference data. We propose SelfRW, a lightweight intrinsic reward that needs no extra fine-tuning or auxiliary models. By pruning current LLMs to approximate an “trust” and an “untrust” token distribution, we compute the log-probability difference as an auxiliary reward. When integrated into reward-guided sampling, SelfRW significantly reduces untrustworthy outputs while preserving task performance. It also improves reward-guided data selection, yielding better post-trained models. Experiments with two reward models and four LLMs on privacy, bias, and stereotype benchmarks show that combining SelfRW consistently improves trustworthiness (over 10% in privacy tasks and 20% in bias tasks) with minimal impact on general utility benchmarks.
Distilling the Essence, Discarding the Dross: Improving Fairness in Multimodal Large Language Models via Historical Reflection-Guided Prompt Optimization
PDF ↗ocial bias in Multimodal Large Language Models (MLLMs) has become an increasingly important concern. Prompt-based approaches offer a lightweight solution for debiasing; however, existing methods rely heavily on handcrafted prompts that are brittle, highly context-sensitive, and difficult to generalize across tasks, bias types, and multimodal settings. In this work, we propose Historical Reflection-Guided Prompt Optimization (HRPO), an adaptive self-debiasing framework for black-box MLLMs that automatically optimizes task-specific debiasing prompts to suppress stereotypical outputs. To mitigate forgetting during prompt optimization, we introduce Historical Contrastive Self-Reflection (HCSR), which performs contrastive reflection over positive and negative optimization histories, enabling the model to retain effective prompts and avoid redundant exploration, thereby improving optimization efficiency. Experiments on three benchmarks involving eight open-source and two closed-source MLLMs, covering ten singular and two intersectional bias types, demonstrate that HRPO achieves strong debiasing performance while offering improved interpretability, generalization, and robustness. Code is available at: https://github.com/liyingji1996/HRPO.
Large Language Models (LLMs) exhibit social biases, which can lead to harmful stereotypes and unfair outcomes. We propose Multi-Persona Thinking (MPT), a simple inference-time framework that reduces social bias by encouraging reasoning from multiple perspectives. MPT guides the model to consider contrasting social identities, such as male and female, together with a neutral viewpoint. These viewpoints then interact through an iterative reasoning process to identify and correct biased judgments. This design transforms the potential weakness of persona assignment into a mechanism to mitigate bias. We evaluate MPT on two widely used bias benchmarks with both open-source and closed-source models. Our results show that MPT achieves a lower bias than the existing prompting-based methods while maintaining the core reasoning ability.
The transition to end-to-end Multimodal Large Language Models (MLLMs) has positioned these architectures as active social evaluators in high-stakes domains. However, it remains unclear whether these models maintain objective auditory perception or succumb to the "Hearing with Eyes" phenomenon, where visual racial cues distort linguistic proficiency evaluations. We investigate this cross-modal bias by constructing a controlled counterfactual dataset utilizing a Visual Matched-Guise Paradigm. By pairing identical native audio with diverse visual personas across English and Korean contexts, we reveal a distinct Cultural Asymmetry in model behavior. In Anglophone settings, most closed models exhibit Reverse Linguistic Stereotyping, hallucinating non-native accents for Asian speakers despite standard native audio. Conversely, in Korean settings, the same models assign baseline-relative competence premiums across all visual personas, with the largest gains for out-group (White/Black) speakers, consistent with Expectancy Violation Theory. Our findings demonstrate that MLLMs do not merely process sensory inputs but actively reproduce context-dependent sociolinguistic ideologies.
Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position preferences). However, prior paradigms typically address these in isolation, often mitigating one at the expense of exacerbating the other. To address this, we conduct a systematic exploration of these reasoning failures and identify a primary inducement: the latent spurious feature correlations within the input that drive these erroneous reasoning shortcuts. Driven by these findings, we introduce Causal-Contrastive Preference Optimization (C2PO), a unified alignment framework designed to tackle these specific failures by simultaneously discovering and suppressing these correlations directly within the optimization process. Specifically, C2PO leverages causal counterfactual signals to isolate bias-inducing features from valid reasoning paths, and employs a fairness-sensitive preference update mechanism to dynamically evaluate logit-level contributions and suppress shortcut features. Extensive experiments across multiple benchmarks covering stereotypical bias (BBQ, Unqover), structural bias (MNLI, HANS, Chatbot, MT-Bench), out-of-domain fairness (StereoSet, WinoBias), and general utility (MMLU, GSM8K) demonstrate that C2PO effectively mitigates stereotypical and structural biases while preserving robust general reasoning capabilities.
While recent safety guardrails effectively suppress overtly biased outputs, subtler forms of social bias emerge during complex logical reasoning tasks that evade current evaluation benchmarks. To fill this gap, we introduce a new evaluation framework, PRIME (\textbf{P}uzzle \textbf{R}easoning for \textbf{I}mplicit Biases in \textbf{M}odel \textbf{E}valuation), that uses logic grid puzzles to systematically probe the influence of social stereotypes on logical reasoning and decision making in LLMs. Our use of logic puzzles enables automatic generation and verification, as well as variability in complexity and biased settings. PRIME includes stereotypical, anti-stereotypical, and neutral puzzle variants generated from a shared puzzle structure, allowing for controlled and fine-grained comparisons. We evaluate multiple model families across puzzle sizes and test the effectiveness of prompt-based mitigation strategies. Focusing our experiments on gender stereotypes, our findings highlight that models consistently reason more accurately when solutions align with stereotypical associations. This demonstrates the significance of PRIME for diagnosing and quantifying social biases perpetuated in the deductive reasoning of LLMs, where fairness is critical.
When a student fails an exam, do we tend to blame their effort or the test’s difficulty? Attribution, defined as how reasons are assigned to event outcomes, shapes perceptions, reinforces stereotypes, and influences decisions. Attribution Theory explains how people attribute causes to internal factors (effort, ability) or external ones (task difficulty, luck). LLMs’ attribution of event outcomes based on demographics carries important fairness implications. Most works exploring social biases in LLMs focus on surface-level associations or isolated stereotypes. This work proposes a cognitively grounded bias evaluation framework to identify how models’ reasoning disparities shape demographic bias across three contexts: single-actor, actor–actor, and actor–observer, capturing comparative and perspective-driven biases overlooked in prior work. Introducing a 140k-prompt benchmark covering ten scenarios and four social dimensions, our analyses reveal attribution asymmetries across identities that vary in multi-actor and observer settings, suggesting that other identities influence bias. This work underscores the need for cognitively grounded bias evaluation and informs future debiasing efforts through the proposed framework.
When Debiasing Backfires: Counterintuitive Side Effects of Preprocessing-Based Stereotype Mitigation
PDF ↗Preprocessing-based methods for stereotype mitigation, such as pre-/post-training on debiased corpora, are widely used in NLP. While these approaches reduce measurable stereotypes for targeted groups, we find they often induce unintended shifts: stereotyping or counter-stereotyping can increase for other demographics, including across unrelated categories. We demonstrate these side effects across two model families (encoder-only and decoder-only), multiple preprocessing strategies (removing stereotypical sentences, removing group mentions, and swapping references), and both pre- and post-training at different data scales on Wikipedia. Standard benchmarks frequently miss these shifts. Using attention-rollout analysis, we observe that such side effects are not accompanied by large changes in attention flow, complicating mechanistic explanations. We discuss implications for evaluation, provide actionable diagnostics, and argue for side-effect-aware, transparent mitigation practices that make claims calibrated to uncertainty.
Large language models (LLMs) are increasingly capable of generating personalized, persuasive text at scale, raising new questions about bias and fairness in automated communication. This paper presents the first systematic analysis of how LLMs behave when tasked with demographic-conditioned targeted messaging. We introduce a controlled evaluation framework using three leading models: GPT-4o, Llama-3.3, and Mistral-Large-2.1, across two generation settings: **Standalone Generation**, which isolates intrinsic demographic effects, and **Context-Rich Generation**, which incorporates thematic and regional context to emulate realistic targeting. We evaluate generated messages along three dimensions: *lexical content*, *language style*, and *persuasive framing*. We instantiate this framework on ***climate communication*** and find consistent age- and gender-based asymmetries across models: male- and youth-targeted messages tend to emphasize more *assertive* and *progressive* framing, while female- and senior-targeted messages more often reflect *warmth*, *care*, and *traditional* themes. Contextual prompts systematically amplify these disparities, with *persuasion* scores being higher for male-targeted messages, while age-related differences vary across models. Our findings demonstrate how demographic stereotypes can surface and intensify in LLM-generated targeted communication, underscoring the need for bias-aware generation pipelines and transparent auditing frameworks that explicitly account for demographic conditioning in socially sensitive applications.
As low-resourced languages are increasingly incorporated into NLP research, there is an emphasis on collecting large-scale datasets. But in prioritizing quantity over quality, we risk 1) building language technologies that perform poorly for these languages and 2) producing harmful content that perpetuates societal biases. In this paper, we investigate the quality of Machine Translation (MT) datasets for three low-resourced languages–Afan Oromo, Amharic, and Tigrinya, with a focus on the gender representation in the datasets. Our findings demonstrate that while training data has a large representation of political and religious domain text, benchmark datasets are focused on news, health, and sports. We also found a large skew towards the male gender–in names of persons, the grammatical gender of verbs, and in stereotypical depictions in the datasets. Further, we found harmful and toxic depictions against women, which were more prominent for the language with the largest amount of data, underscoring that quantity does not guarantee quality. We hope that our work inspires further inquiry into the datasets collected for low-resourced languages and prompts early mitigation of harmful content. WARNING: This paper contains discussion of NSFW content that some may find disturbing.
Large language models (LLMs) are known to exhibit gender bias, yet most evaluations focus on downstream stereotypes rather than the normative frameworks that shape model inference. We investigate whether LLMs rely on traditional masculinity norms (e.g. "real men are tough") as latent priors in gender-biased inference. We ground our evaluation in the Male Role Norms Inventory (MRNI), a validated psychological framework of prescriptive male role norms.Anchored in MRNI items, we probe models using two complementary approaches: (i) explicit Likert-style agreement with masculinity norms, and (ii) a newly crafted English-Italian scenario-based inference dataset (MRNI-BB), in which gender information and evidential support are systematically varied. Across models, explicit endorsement of masculinity norms is generally low. In contrast, in scenario-based inference tasks, models systematically attribute MRNI-aligned behaviors to male agents, even when evidence is ambiguous or absent. This effect disappears when gender markers are removed, suggesting that masculinity norms are treated as gender-specific expectations about male agents. Increasing model scale reduces explicit norm endorsement but is associated with stronger male-directed bias under uncertainty.
Adaptive Data Collection for Latin-American Community-sourced Evaluation of Stereotypes (LACES)
PDF ↗The evaluation of societal biases in NLP models is critically hindered by a geo-cultural gap. This leaves regions such as Latin America severely underserved, making it impossible to adequately assess or mitigate the perpetuation of harmful regional stereotypes in language technologies. This paper presents LACES, a stereotype association dataset, for 15 Latin American countries. This dataset includes 4,789 stereotype associations[The de-identified dataset can be accessed via GitHub], manually created and annotated by 83 participants. The dataset was developed through targeted community partnerships across Latin America. Additionally, in this paper, we propose a novel adaptive data collection methodology that uniquely integrates the sourcing of new stereotype entries and the validation of existing data within a single, unified workflow. This approach results in a resource with more unique stereotypes than previous static collection methods, enabling a more efficient stereotype collection. The paper further supports the quality of LACES by demonstrating reduced efficacy of debiasing methods on this dataset in comparison to existing popular stereotype benchmarks.Content Warning: This research involves the study of social biases. Consequently, the paper contains examples of discriminatory language and stereotypes that may be sensitive or upsetting to readers. These examples are included for the purpose of scientific analysis and do not reflect the views of the authors.
Rethinking Research on Stereotypes: An Analysis through Social Psychological and Computational Perspectives
PDF ↗Stereotypes are social constructs shaping human perception and behavior that can produce harmful outcomes under specific conditions. Recent work shows that large language models (LLMs) may inherit and amplify such social harms. However, most existing research often focuses only upon stereotypical biases and overlooks stereotypes and the rich social psychological literature on them, resulting in resource wastage and slowed progress in stereotype research.We argue that meaningful progress in mitigating stereotypes in LLMs requires tighter integration between social psychology and computational research. To address this gap, we review core social psychological theories and frameworks and analyze their computational operationalization, highlighting substantial open opportunities.We also analyze computational progress across media narratives, body imaging, and multilingual, multicultural, and multimodal contexts, identifying key gaps and limitations in each domain.We also present a unified analysis of challenges in stereotype research.We further discuss implications for responsible AI, highlighting stereotypes as a major source of downstream harms, and briefly examine the limitations of current mitigation approaches along with potential improvements via explainability and interpretability. We frame stereotypes in AI as socio-technical phenomena and urge further research in responsible AI, informed by the perspectives and future directions presented in this paper.
Large Language Models (LLMs) can learn both useful knowledge and harmful stereotypes, making bias evaluation essential.Existing frameworks fall into two types: those considering reasoning steps (Thinking Process-Aware Evaluation, TPAE) and those focusing only on final outputs (Straight-to-the-Answer Evaluation, SAE).Prior TPAE studies showed effectiveness in assessing gender bias but relied on template-based, word-counting prompts, limiting generalization to other bias types, languages, and reasoning-based methods.In this study, we introduce MBTP, a multilingual social bias benchmark that incorporates human-generated pro- and anti-stereotype reasoning as part of the thinking process, and propose a few-shot meta-evaluation method that enables scalable bias assessment without model fine-tuning.From experiments evaluating 13 social bias categories across 8 languages, we find that human-generated thinking consistently yields higher-quality evaluations than LLM-generated or template-based approaches.Furthermore, TPAE demonstrates superior performance over SAE, highlighting the importance of considering reasoning processes in bias evaluation.We will release the MBTP dataset upon paper acceptance.
Humor holds up a mirror to social perception: what we find funny often reflects who we are and how we judge others. When language models engage with humor, their reactions expose the social assumptions they have internalized from training data. In this paper, we investigate counterfactual unfairness through humor by observing how the model’s responses change when we swap who speaks and who is addressed while holding other factors constant. Our framework spans three tasks: humor generation refusal, speaker intention inference, and relational/societal impact prediction, covering both identity-agnostic humor and identity-specific disparagement humor. We introduce interpretable bias metrics that capture asymmetric patterns under identity swaps. Experiments across state-of-the-art models reveal consistent relational disparities: jokes told by privileged speakers are refused up to 67.5% more often, judged as malicious 64.7% more frequently, and rated up to 1.5 points higher in social harm on a 5-point scale. These patterns highlight how sensitivity and stereotyping coexist in generative models, complicating efforts toward fairness and cultural alignment.
Cultural alignment in Large Language Models (LLMs) is essential for producing contextually aware, respectful, and trustworthy outputs. Without it, models risk generating stereotyped, insensitive, or misleading responses that fail to reflect cultural diversity w.r.t Helpful, Harmless, and Honest (HHH) paradigm. Existing benchmarks represent early steps toward cultural alignment; yet, no benchmarks currently enables systematic evaluation of cultural alignment in line with UNESCO’s principles of cultural diversity w.r.t HHH paradigm. Therefore, to address this gap, we built Align-Cultura, two-stage pipeline for cultural alignment. Stage I constructs CULTURAX, the HHH-English dataset grounded in the UNESCO cultural taxonomy, through Query Construction, which reclassifies prompts, expands underrepresented domains (or labels), and prevents data leakage with SimHash. Then, Response Generation pairs prompts with culturally grounded responses via two-stage rejection sampling. The final dataset contains 1,500 samples spanning 30 subdomains of tangible and intangible cultural forms. Stage II benchmarks CULTURAX on general-purpose models, culturally fine-tuned models, and open-weight LLMs (Qwen3-8B and DeepSeek-R1-Distill-Qwen-7B). Empirically, culturally fine-tuned models improve joint HHH by 4%–6%, reduce cultural failures by 18%, achieve 10%–12% efficiency gains, and limit leakage to 0.3%.
Persona-E²: A Human-Grounded Dataset for Personality-Shaped Emotional Responses to Textual Events
PDF ↗Most affective computing research treats emotion as a static property of text, focusing on the writer’s sentiment while overlooking the reader’s perspective. This approach ignores how individual personalities lead to diverse emotional appraisals of the same event. Although role-playing Large Language Models (LLMs) attempt to simulate such nuanced reactions, they often suffer from "personality illusion”—relying on surface-level stereotypes rather than authentic cognitive logic. A critical bottleneck is the absence of ground-truth human data to link personality traits to emotional shifts. To bridge the gap, we introduce Persona-E² (Persona-Event2Emotion), a large-scale dataset grounded in annotated MBTI and Big Five traits to capture reader-based emotional variations across news, social media, and life narratives. Extensive experiments reveal that state-of-the-art LLMs struggle to capture precise appraisal shifts, particularly in social media domains. Crucially, we find that personality information significantly improves comprehension, with the Big Five traits alleviating "personality illusion.’
Vision language models (VLMs) demonstrate strong zero-shot performance, but often perpetuate social stereotypes in person-centric queries, yielding skewed demographic distributions. Current debiasing methods apply uniform bias corrections across all input queries regardless of their bias sensitivity, creating a fundamental fairness–utility trade-off. Strong debiasing distorts semantically meaningful information in bias-insensitive queries, while weak debiasing fails to mitigate stereotypes in bias-sensitive ones. This one-size-fits-all approach hampers simultaneously achieving high utility on bias-insensitive queries and fairness on bias-sensitive queries. We introduce Reward-Gated Test-Time Adaptation (RG-TTA), a reinforcement learning-based test-time adaptation framework that selectively applies debiasing based on input sensitivity. RG-TTA adaptively triggers fairness regularization based on the bias sensitivity of each input during test-time policy adaptation, while focusing exclusively on optimizing cross-modal alignment for bias-insensitive inputs. Experiments on fairness benchmarks (e.g., FairFace, UTKFace) demonstrate substantial bias reduction while simultaneously improving zero-shot utility, resolving the trade-off of uniform debiasing.