Kristina Lerman

Georgios Chochlakis, Gireesh Mahajan, Sabyasachee Baruah et al.

Detecting emotions expressed in text has become critical to a range of fields. In this work, we investigate ways to exploit label correlations in multi-label emotion recognition models to improve emotion detection. First, we develop two modeling approaches to the problem in order to capture word associations of the emotion words themselves, by either including the emotions in the input, or by leveraging Masked Language Modeling (MLM). Second, we integrate pairwise constraints of emotion representations as regularization terms alongside the classification loss of the models. We split these terms into two categories, local and global. The former dynamically change based on the gold labels, while the latter remain static during training. We demonstrate state-of-the-art performance across Spanish, English, and Arabic in SemEval 2018 Task 1 E-c using monolingual BERT-based models. On top of better performance, we also demonstrate improved robustness. Code is available at https://github.com/gchochla/Demux-MEmo.

Georgios Chochlakis, Gireesh Mahajan, Sabyasachee Baruah et al.

The need for emotional inference from text continues to diversify as more and more disciplines integrate emotions into their theories and applications. These needs include inferring different emotion types, handling multiple languages, and different annotation formats. A shared model between different configurations would enable the sharing of knowledge and a decrease in training costs, and would simplify the process of deploying emotion recognition models in novel environments. In this work, we study how we can build a single model that can transition between these different configurations by leveraging multilingual models and Demux, a transformer-based model whose input includes the emotions of interest, enabling us to dynamically change the emotions predicted by the model. Demux also produces emotion embeddings, and performing operations on them allows us to transition to clusters of emotions by pooling the embeddings of each cluster. We show that Demux can simultaneously transfer knowledge in a zero-shot manner to a new language, to a novel annotation format and to unseen emotions. Code is available at https://github.com/gchochla/Demux-MEmo .

Negar Mokhberian, Myrl G. Marmarelis, Frederic R. Hopp et al.

Supervised classification heavily depends on datasets annotated by humans. However, in subjective tasks such as toxicity classification, these annotations often exhibit low agreement among raters. Annotations have commonly been aggregated by employing methods like majority voting to determine a single ground truth label. In subjective tasks, aggregating labels will result in biased labeling and, consequently, biased models that can overlook minority opinions. Previous studies have shed light on the pitfalls of label aggregation and have introduced a handful of practical approaches to tackle this issue. Recently proposed multi-annotator models, which predict labels individually per annotator, are vulnerable to under-determination for annotators with few samples. This problem is exacerbated in crowdsourced datasets. In this work, we propose \textbf{Annotator Aware Representations for Texts (AART)} for subjective classification tasks. Our approach involves learning representations of annotators, allowing for exploration of annotation behaviors. We show the improvement of our method on metrics that assess the performance on capturing individual annotators' perspectives. Additionally, we demonstrate fairness metrics to evaluate our model's equability of performance for marginalized annotators compared to others.

Georgios Chochlakis, Niyantha Maruthu Pandiyan, Kristina Lerman et al.

In-Context Learning (ICL) in Large Language Models (LLM) has emerged as the dominant technique for performing natural language tasks, as it does not require updating the model parameters with gradient-based methods. ICL promises to "adapt" the LLM to perform the present task at a competitive or state-of-the-art level at a fraction of the computational cost. ICL can be augmented by incorporating the reasoning process to arrive at the final label explicitly in the prompt, a technique called Chain-of-Thought (CoT) prompting. However, recent work has found that ICL relies mostly on the retrieval of task priors and less so on "learning" to perform tasks, especially for complex subjective domains like emotion and morality, where priors ossify posterior predictions. In this work, we examine whether "enabling" reasoning also creates the same behavior in LLMs, wherein the format of CoT retrieves reasoning priors that remain relatively unchanged despite the evidence in the prompt. We find that, surprisingly, CoT indeed suffers from the same posterior collapse as ICL for larger language models. Code is avalaible at https://github.com/gchochla/cot-priors.

Wanying Zhao, Siyi Guo, Kristina Lerman et al.

Narrative is a foundation of human cognition and decision making. Because narratives play a crucial role in societal discourses and spread of misinformation and because of the pervasive use of social media, the narrative dynamics on social media can have profound societal impact. Yet, systematic and computational understanding of online narratives faces critical challenge of the scale and dynamics; how can we reliably and automatically extract narratives from massive amount of texts? How do narratives emerge, spread, and die? Here, we propose a systematic narrative discovery framework that fill this gap by combining change point detection, semantic role labeling (SRL), and automatic aggregation of narrative fragments into narrative networks. We evaluate our model with synthetic and empirical data two-Twitter corpora about COVID-19 and 2017 French Election. Results demonstrate that our approach can recover major narrative shifts that correspond to the major events.

Negar Mokhberian, Frederic R. Hopp, Bahareh Harandizadeh et al.

Morality plays an important role in culture, identity, and emotion. Recent advances in natural language processing have shown that it is possible to classify moral values expressed in text at scale. Morality classification relies on human annotators to label the moral expressions in text, which provides training data to achieve state-of-the-art performance. However, these annotations are inherently subjective and some of the instances are hard to classify, resulting in noisy annotations due to error or lack of agreement. The presence of noise in training data harms the classifier's ability to accurately recognize moral foundations from text. We propose two metrics to audit the noise of annotations. The first metric is entropy of instance labels, which is a proxy measure of annotator disagreement about how the instance should be labeled. The second metric is the silhouette coefficient of a label assigned by an annotator to an instance. This metric leverages the idea that instances with the same label should have similar latent representations, and deviations from collective judgments are indicative of errors. Our experiments on three widely used moral foundations datasets show that removing noisy annotations based on the proposed metrics improves classification performance.

Zihao He, Negar Mokhberian, Kristina Lerman

Stance detection infers a text author's attitude towards a target. This is challenging when the model lacks background knowledge about the target. Here, we show how background knowledge from Wikipedia can help enhance the performance on stance detection. We introduce Wikipedia Stance Detection BERT (WS-BERT) that infuses the knowledge into stance encoding. Extensive results on three benchmark datasets covering social media discussions and online debates indicate that our model significantly outperforms the state-of-the-art methods on target-specific stance detection, cross-target stance detection, and zero/few-shot stance detection.

Daniel Hickey, Matheus Schmitz, Daniel M. T. Fessler et al.

Counterspeech has gained attention as a strategy to reduce hate speech on social media. Although previous studies suggest that counterspeech can reduce hate speech, little is known about its effects on participation in online hate communities. Relatedly, we lack an understanding about the degree of hostility in counterspeech. Hostile counterspeech may increase online conflict, potentially hardening the positions of hate adherents, and further eroding online environments. Here, we analyzed the effect of counterspeech on 16,513 newcomers across 104 hate subreddits (forums within Reddit.com). We devised an LLM-based counterspeech detection approach that outperforms specialized models trained on existing datasets, then examined the presence, and effects of, hostility. While counterspeech comments are less toxic than hate speech comments, they are almost twice as toxic as other discourse within hate subreddits. We then evaluated the effect of counterspeech on newcomer engagement in hate subreddits. We found that newcomers using hate speech who receive counterspeech are less likely to continue posting within these hate subreddits, rather than becoming galvanized. We speculate that, instead of constituting ardent hate adherents, readily-dissuaded newcomers may merely be toying with beliefs that are proscribed in other contexts. Although we found no association between the toxicity of counterspeech and its effects on user retention, consistent with prior research regarding the harmful effects of toxic speech, we found that toxic counterspeech increases the probability of continued hostility from hate users within the same discussion.

Kai Chen, Zihao He, Rong-Ching Chang et al.

Emotions play an important role in interpersonal interactions and social conflict, yet their function in the development of controversy and disagreement in online conversations has not been explored. To address this gap, we study controversy on Reddit, a popular network of online discussion forums. We collect discussions from a wide variety of topical forums and use emotion detection to recognize a range of emotions from text, including anger, fear, joy, admiration, etc. Our study has three main findings. First, controversial comments express more anger and less admiration, joy and optimism than non-controversial comments. Second, controversial comments affect emotions of downstream comments in a discussion, usually resulting in long-term increase in anger and a decrease in positive emotions, although the magnitude and direction of emotional change depends on the forum. Finally, we show that emotions help better predict which comments will become controversial. Understanding emotional dynamics of online discussions can help communities to better manage conversations.

Siyi Guo, Negar Mokhberian, Kristina Lerman

Language models can be trained to recognize the moral sentiment of text, creating new opportunities to study the role of morality in human life. As interest in language and morality has grown, several ground truth datasets with moral annotations have been released. However, these datasets vary in the method of data collection, domain, topics, instructions for annotators, etc. Simply aggregating such heterogeneous datasets during training can yield models that fail to generalize well. We describe a data fusion framework for training on multiple heterogeneous datasets that improve performance and generalizability. The model uses domain adversarial training to align the datasets in feature space and a weighted loss function to deal with label shift. We show that the proposed framework achieves state-of-the-art performance in different datasets compared to prior works in morality inference.

Zihao He, Weituo Hao, Wei-Tsung Lu et al.

Music captioning has gained significant attention in the wake of the rising prominence of streaming media platforms. Traditional approaches often prioritize either the audio or lyrics aspect of the music, inadvertently ignoring the intricate interplay between the two. However, a comprehensive understanding of music necessitates the integration of both these elements. In this study, we delve into this overlooked realm by introducing a method to systematically learn multimodal alignment between audio and lyrics through contrastive learning. This not only recognizes and emphasizes the synergy between audio and lyrics but also paves the way for models to achieve deeper cross-modal coherence, thereby producing high-quality captions. We provide both theoretical and empirical results demonstrating the advantage of the proposed method, which achieves new state-of-the-art on two music captioning datasets.

Zihao He, Siyi Guo, Ashwin Rao et al.

Social media platforms are rife with politically charged discussions. Therefore, accurately deciphering and predicting partisan biases using Large Language Models (LLMs) is increasingly critical. In this study, we address the challenge of understanding political bias in digitized discourse using LLMs. While traditional approaches often rely on finetuning separate models for each political faction, our work innovates by employing a singular, instruction-tuned LLM to reflect a spectrum of political ideologies. We present a comprehensive analytical framework, consisting of Partisan Bias Divergence Assessment and Partisan Class Tendency Prediction, to evaluate the model's alignment with real-world political ideologies in terms of stances, emotions, and moral foundations. Our findings reveal the model's effectiveness in capturing emotional and moral nuances, albeit with some challenges in stance detection, highlighting the intricacies and potential for refinement in NLP tools for politically sensitive contexts. This research contributes significantly to the field by demonstrating the feasibility and importance of nuanced political understanding in LLMs, particularly for applications requiring acute awareness of political bias.

Minh Duc Chu, Zihao He, Rebecca Dorn et al.

Large language models (LLMs) have shown promise in representing individuals and communities, offering new ways to study complex social dynamics. However, effectively aligning LLMs with specific human groups and systematically assessing the fidelity of the alignment remains a challenge. This paper presents a robust framework for aligning LLMs with online communities via instruction-tuning and comprehensively evaluating alignment across various aspects of language, including authenticity, emotional tone, toxicity, and harm. We demonstrate the utility of our approach by applying it to online communities centered on dieting and body image. We administer an eating disorder psychometric test to the aligned LLMs to reveal unhealthy beliefs and successfully differentiate communities with varying levels of eating disorder risk. Our results highlight the potential of LLMs in automated moderation and broader applications in public health and social science research.

Christopher Tran, Keith Burghardt, Kristina Lerman et al.

Estimating how a treatment affects different individuals, known as heterogeneous treatment effect estimation, is an important problem in empirical sciences. In the last few years, there has been a considerable interest in adapting machine learning algorithms to the problem of estimating heterogeneous effects from observational and experimental data. However, these algorithms often make strong assumptions about the observed features in the data and ignore the structure of the underlying causal model, which can lead to biased estimation. At the same time, the underlying causal mechanism is rarely known in real-world datasets, making it hard to take it into consideration. In this work, we provide a survey of state-of-the-art data-driven methods for heterogeneous treatment effect estimation using machine learning, broadly categorizing them as methods that focus on counterfactual prediction and methods that directly estimate the causal effect. We also provide an overview of a third category of methods which rely on structural causal models and learn the model structure from data. Our empirical evaluation under various underlying structural model mechanisms shows the advantages and deficiencies of existing estimators and of the metrics for measuring their performance.

Minh Duc Chu, Cinthia Sánchez, Zihao He et al.

The pervasive expectations about ideal body types in Western society can lead to body image concerns, dissatisfaction, and in extreme cases, eating disorders and other psychopathologies related to body image. While previous research has focused on online pro-anorexia communities glorifying the "thin ideal," less attention has been given to the broader spectrum of body image concerns or how emerging disorders like muscle dysmorphia ("bigorexia") present on online platforms. To address this gap, we analyze 46 Reddit forums related to diet, fitness, and mental health. We map these communities along gender and body ideal dimensions, revealing distinct patterns of emotional expression and community support. Feminine-oriented communities, especially those endorsing the thin ideal, express higher levels of negative emotions and receive caring comments in response. In contrast, muscular ideal communities display less negativity, regardless of gender orientation, but receive aggressive compliments in response, marked by admiration and toxicity. Mental health discussions align more with thin ideal, feminine-leaning spaces. By uncovering these gendered emotional dynamics, our findings can inform the development of moderation strategies that foster supportive interactions while reducing exposure to harmful content.

Ashwin Rao, Rong-Ching Chang, Qiankun Zhong et al.

Abortion has been one of the most divisive issues in the United States. Yet, missing is comprehensive longitudinal evidence on how political divides on abortion are reflected in public discourse over time, on a national scale, and in response to key events before and after the overturn of Roe v Wade. We analyze a corpus of over 3.5M tweets related to abortion over the span of one year (January 2022 to January 2023) from over 1.1M users. We estimate users' ideology and rely on state-of-the-art transformer-based classifiers to identify expressions of hostility and extract five prominent frames surrounding abortion. We use those data to examine (a) how prevalent were expressions of hostility (i.e., anger, toxic speech, insults, obscenities, and hate speech), (b) what frames liberals and conservatives used to articulate their positions on abortion, and (c) the prevalence of hostile expressions in liberals and conservative discussions of these frames. We show that liberals and conservatives largely mirrored each other's use of hostile expressions: as liberals used more hostile rhetoric, so did conservatives, especially in response to key events. In addition, the two groups used distinct frames and discussed them in vastly distinct contexts, suggesting that liberals and conservatives have differing perspectives on abortion. Lastly, frames favored by one side provoked hostile reactions from the other: liberals use more hostile expressions when addressing religion, fetal personhood, and exceptions to abortion bans, whereas conservatives use more hostile language when addressing bodily autonomy and women's health. This signals disrespect and derogation, which may further preclude understanding and exacerbate polarization.

Kristina Lerman

Algorithmic systems, particularly social media recommenders, have achieved remarkable success in predicting behavior. By optimizing for observable signals such as clicks, views, and engagement, these systems effectively capture user attention and guide interaction. Yet their widespread adoption has coincided with troubling outcomes, including rising mental health concerns, increasing polarization, and erosion of trust. This paper argues that these effects are consequences of a structural functional misalignment between what algorithms optimize - predictable behavior - and the human goals these predictions are intended to serve. We propose that this misalignment arises through three mechanisms: (1) a bias toward modeling fast, reactive behavioral signals over reflective judgment, (2) feedback loops that couple user behavior with algorithmic learning, and (3) emergent collective dynamics that amplify these effects at scale. Together, these mechanisms explain how accurate individual-level predictions can produce adverse societal outcomes. We present functional misalignment as a unifying framework and outline a research agenda for studying and mitigating its effects in human-AI interaction systems.

Louis Penafiel, Hsien-Te Kao, Isabel Erickson et al.

Eating disorders are complex mental health conditions that affect millions of people around the world. Effective interventions on social media platforms are crucial, yet testing strategies in situ can be risky. We present a novel LLM-driven experimental testbed for simulating and assessing intervention strategies in ED-related discussions. Our framework generates synthetic conversations across multiple platforms, models, and ED-related topics, allowing for controlled experimentation with diverse intervention approaches. We analyze the impact of various intervention strategies on conversation dynamics across four dimensions: intervention type, generative model, social media platform, and ED-related community/topic. We employ cognitive domain analysis metrics, including sentiment, emotions, etc., to evaluate the effectiveness of interventions. Our findings reveal that civility-focused interventions consistently improve positive sentiment and emotional tone across all dimensions, while insight-resetting approaches tend to increase negative emotions. We also uncover significant biases in LLM-generated conversations, with cognitive metrics varying notably between models (Claude-3 Haiku $>$ Mistral $>$ GPT-3.5-turbo $>$ LLaMA3) and even between versions of the same model. These variations highlight the importance of model selection in simulating realistic discussions related to ED. Our work provides valuable information on the complex dynamics of ED-related discussions and the effectiveness of various intervention strategies.

Patrick Gerard, William Theisen, Tim Weninger et al.

Othering, the act of portraying outgroups as fundamentally different from the ingroup, often escalates into framing them as existential threats--fueling intergroup conflict and justifying exclusion and violence. These dynamics are alarmingly pervasive, spanning from the extreme historical examples of genocides against minorities in Germany and Rwanda to the ongoing violence and rhetoric targeting migrants in the US and Europe. While concepts like hate speech and fear speech have been explored in existing literature, they capture only part of this broader and more nuanced dynamic which can often be harder to detect, particularly in online speech and propaganda. To address this challenge, we introduce a novel computational framework that leverages large language models (LLMs) to quantify othering across diverse contexts, extending beyond traditional linguistic indicators of hostility. Applying the model to real-world data from Telegram war bloggers and political discussions on Gab reveals how othering escalates during conflicts, interacts with moral language, and garners significant attention, particularly during periods of crisis. Our framework, designed to offer deeper insights into othering dynamics, combines with a rapid adaptation process to provide essential tools for mitigating othering's adverse impacts on social cohesion.

Georgios Chochlakis, Peter Wu, Arjun Bedi et al.

Modeling complex subjective tasks in Natural Language Processing, such as recognizing emotion and morality, is considerably challenging due to significant variation in human annotations. This variation often reflects reasonable differences in semantic interpretations rather than mere noise, necessitating methods to distinguish between legitimate subjectivity and error. We address this challenge by exploring label verification in these contexts using Large Language Models (LLMs). First, we propose a simple In-Context Learning binary filtering baseline that estimates the reasonableness of a document-label pair. We then introduce the Label-in-a-Haystack setting: the query and its label(s) are included in the demonstrations shown to LLMs, which are prompted to predict the label(s) again, while receiving task-specific instructions (e.g., emotion recognition) rather than label copying. We show how the failure to copy the label(s) to the output of the LLM are task-relevant and informative. Building on this, we propose the Label-in-a-Haystack Rectification (LiaHR) framework for subjective label correction: when the model outputs diverge from the reference gold labels, we assign the generated labels to the example instead of discarding it. This approach can be integrated into annotation pipelines to enhance signal-to-noise ratios. Comprehensive analyses, human evaluations, and ecological validity studies verify the utility of LiaHR for label correction. Code is available at https://github.com/gchochla/liahr.

Charles Bickham, Minh Duc Chu, Arianna Yuan et al.

Whether genuine communities can form on algorithmically-driven short-form video platforms like TikTok remains an open question, given that user interactions are often brief, dispersed, and difficult to trace. Building on theories of tie strength and online community formation, we examine whether eating disorder (ED) discourse on TikTok exhibits behavioral and emotional signatures of strong ties, including more frequent, reciprocal, and affectively intense interactions. In this paper, we analyze 43,040 ED-related TikTok videos and over 560,000 comments, alongside a Non-ED comparison dataset. We find that at the user-pair level, greater interaction frequency is associated with increasingly positive emotional expression, a pattern that is amplified in ED-related conversations. This trend is also reflected linguistically, with pairs that interact more frequently exhibiting more of a positive tone. At the same time, how a relationship starts matters: pairs that begin with positive exchanges usually stay mostly positive as they continue interacting, while pairs that begin negatively may add some positive exchanges over time but rarely become mostly positive. To contextualize these dynamics, we classify ED videos into three content types (Pro-Recovery, Pro-ED, and ED Experiences) and find that each exhibits distinct emotional interaction patterns. These findings suggest that dense, emotionally structured relationships can emerge within ED discourse on TikTok. More broadly, our work provides one of the first empirical demonstrations of how community-like relational dynamics form and persist on algorithmically driven short-form video platforms.

Rebecca Dorn, Christina Chance, Casandra Rusti et al.

Automated emotion detection is widely used in applications ranging from well-being monitoring to high-stakes domains like mental health and hiring. However, models often rely on annotations that reflect dominant cultural norms, limiting model ability to recognize emotional expression in dialects often excluded from training data distributions, such as African American Vernacular English (AAVE). This study examines emotion recognition model performance on AAVE compared to General American English (GAE). We analyze 2.7 million tweets geo-tagged within Los Angeles. Texts are scored for strength of AAVE using computational approximations of dialect features. Annotations of emotion presence and intensity are collected on a dataset of 875 tweets with both high and low AAVE densities. To assess model accuracy on a task as subjective as emotion perception, we calculate community-informed "silver" labels where AAVE-dense tweets are labeled by African American, AAVE-fluent (ingroup) annotators. On our labeled sample, GPT and BERT-based models exhibit false positive prediction rates of anger on AAVE more than double than on GAE. SpanEmo, a popular text-based emotion model, increases false positive rates of anger from 25 percent on GAE to 60 percent on AAVE. Additionally, a series of linear regressions reveals that models and non-ingroup annotations are significantly more correlated with profanity-based AAVE features than ingroup annotations. Linking Census tract demographics, we observe that neighborhoods with higher proportions of African American residents are associated with higher predictions of anger (Pearson's correlation r = 0.27) and lower joy (r = -0.10). These results find an emergent safety issue of emotion AI reinforcing racial stereotypes through biased emotion classification. We emphasize the need for culturally and dialect-informed affective computing systems.

Patrick Gerard, Svitlana Volkova, Louis Penafiel et al.

Following the Russian Federation's full-scale invasion of Ukraine in February 2022, a multitude of information narratives emerged within both pro-Russian and pro-Ukrainian communities online. As the conflict progresses, so too do the information narratives, constantly adapting and influencing local and global community perceptions and attitudes. This dynamic nature of the evolving information environment (IE) underscores a critical need to fully discern how narratives evolve and affect online communities. Existing research, however, often fails to capture information narrative evolution, overlooking both the fluid nature of narratives and the internal mechanisms that drive their evolution. Recognizing this, we introduce a novel approach designed to both model narrative evolution and uncover the underlying mechanisms driving them. In this work we perform a comparative discourse analysis across communities on Telegram covering the initial three months following the invasion. First, we uncover substantial disparities in narratives and perceptions between pro-Russian and pro-Ukrainian communities. Then, we probe deeper into prevalent narratives of each group, identifying key themes and examining the underlying mechanisms fueling their evolution. Finally, we explore influences and factors that may shape the development and spread of narratives.

Kai Chen, Zihao He, Jun Yan et al.

Large Language Models (LLMs) possess the potential to exert substantial influence on public perceptions and interactions with information. This raises concerns about the societal impact that could arise if the ideologies within these models can be easily manipulated. In this work, we investigate how effectively LLMs can learn and generalize ideological biases from their instruction-tuning data. Our findings reveal a concerning vulnerability: exposure to only a small amount of ideologically driven samples significantly alters the ideology of LLMs. Notably, LLMs demonstrate a startling ability to absorb ideology from one topic and generalize it to even unrelated ones. The ease with which LLMs' ideologies can be skewed underscores the risks associated with intentionally poisoned training data by malicious actors or inadvertently introduced biases by data annotators. It also emphasizes the imperative for robust safeguards to mitigate the influence of ideological manipulations on LLMs.

Zihao He, Siyi Guo, Ashwin Rao et al.

Language models (LMs) are known to represent the perspectives of some social groups better than others, which may impact their performance, especially on subjective tasks such as content moderation and hate speech detection. To explore how LMs represent different perspectives, existing research focused on positional alignment, i.e., how closely the models mimic the opinions and stances of different groups, e.g., liberals or conservatives. However, human communication also encompasses emotional and moral dimensions. We define the problem of affective alignment, which measures how LMs' emotional and moral tone represents those of different groups. By comparing the affect of responses generated by 36 LMs to the affect of Twitter messages, we observe significant misalignment of LMs with both ideological groups. This misalignment is larger than the partisan divide in the U.S. Even after steering the LMs towards specific ideological perspectives, the misalignment and liberal tendencies of the model persist, suggesting a systemic bias within LMs.

Rebecca Dorn, Lee Kezar, Fred Morstatter et al.

Content moderation on social media platforms shapes the dynamics of online discourse, influencing whose voices are amplified and whose are suppressed. Recent studies have raised concerns about the fairness of content moderation practices, particularly for aggressively flagging posts from transgender and non-binary individuals as toxic. In this study, we investigate the presence of bias in harmful speech classification of gender-queer dialect online, focusing specifically on the treatment of reclaimed slurs. We introduce a novel dataset, QueerReclaimLex, based on 109 curated templates exemplifying non-derogatory uses of LGBTQ+ slurs. Dataset instances are scored by gender-queer annotators for potential harm depending on additional context about speaker identity. We systematically evaluate the performance of five off-the-shelf language models in assessing the harm of these texts and explore the effectiveness of chain-of-thought prompting to teach large language models (LLMs) to leverage author identity context. We reveal a tendency for these models to inaccurately flag texts authored by gender-queer individuals as harmful. Strikingly, across all LLMs the performance is poorest for texts that show signs of being written by individuals targeted by the featured slur (F1 <= 0.24). We highlight an urgent need for fairness and inclusivity in content moderation systems. By uncovering these biases, this work aims to inform the development of more equitable content moderation practices and contribute to the creation of inclusive online spaces for all users.

Georgios Chochlakis, Alexandros Potamianos, Kristina Lerman et al.

In-context Learning (ICL) has emerged as a powerful paradigm for performing natural language tasks with Large Language Models (LLM) without updating the models' parameters, in contrast to the traditional gradient-based finetuning. The promise of ICL is that the LLM can adapt to perform the present task at a competitive or state-of-the-art level at a fraction of the cost. The ability of LLMs to perform tasks in this few-shot manner relies on their background knowledge of the task (or task priors). However, recent work has found that, unlike traditional learning, LLMs are unable to fully integrate information from demonstrations that contrast task priors. This can lead to performance saturation at suboptimal levels, especially for subjective tasks such as emotion recognition, where the mapping from text to emotions can differ widely due to variability in human annotations. In this work, we design experiments and propose measurements to explicitly quantify the consistency of proxies of LLM priors and their pull on the posteriors. We show that LLMs have strong yet inconsistent priors in emotion recognition that ossify their predictions. We also find that the larger the model, the stronger these effects become. Our results suggest that caution is needed when using ICL with larger LLMs for affect-centered tasks outside their pre-training domain and when interpreting ICL results.

Abhishek Anand, Negar Mokhberian, Prathyusha Naresh Kumar et al.

Researchers have raised awareness about the harms of aggregating labels especially in subjective tasks that naturally contain disagreements among human annotators. In this work we show that models that are only provided aggregated labels show low confidence on high-disagreement data instances. While previous studies consider such instances as mislabeled, we argue that the reason the high-disagreement text instances have been hard-to-learn is that the conventional aggregated models underperform in extracting useful signals from subjective tasks. Inspired by recent studies demonstrating the effectiveness of learning from raw annotations, we investigate classifying using Multiple Ground Truth (Multi-GT) approaches. Our experiments show an improvement of confidence for the high-disagreement instances.

Hieu Nguyen, Zihao He, Shoumik Atul Gandre et al.

Large language models (LLMs) often suffer from hallucination, generating factually incorrect or ungrounded content, which limits their reliability in high-stakes applications. A key factor contributing to hallucination is the use of hard labels during training, which enforce deterministic supervision, encourage overconfidence, and disregard the uncertainty inherent in natural language. To address this, we propose mitigating hallucination through knowledge distillation (KD), where a teacher model provides smoothed soft labels to a student model, reducing overconfidence and improving factual grounding. We apply KD during supervised finetuning on instructional data, evaluating its effectiveness across LLMs from different families. Experimental results on summarization benchmarks demonstrate that KD reduces hallucination compared to standard finetuning while preserving performance on general NLP tasks. These findings highlight KD as a promising approach for mitigating hallucination in LLMs and improving model reliability.

Zihao He, Ashwin Rao, Siyi Guo et al.

Recent advances in NLP have improved our ability to understand the nuanced worldviews of online communities. Existing research focused on probing ideological stances treats liberals and conservatives as separate groups. However, this fails to account for the nuanced views of the organically formed online communities and the connections between them. In this paper, we study discussions of the 2020 U.S. election on Twitter to identify complex interacting communities. Capitalizing on this interconnectedness, we introduce a novel approach that harnesses message passing when finetuning language models (LMs) to probe the nuanced ideologies of these communities. By comparing the responses generated by LMs and real-world survey results, our method shows higher alignment than existing baselines, highlighting the potential of using LMs in revealing complex ideologies within and across interconnected mixed-ideology communities.

Buddhika Nettasinghe, Ashwin Rao, Bohan Jiang et al.

Affective polarization, the emotional divide between ideological groups marked by in-group love and out-group hate, has intensified in the United States, driving contentious issues like masking and lockdowns during the COVID-19 pandemic. Despite its societal impact, existing models of opinion change fail to account for emotional dynamics nor offer methods to quantify affective polarization robustly and in real-time. In this paper, we introduce a discrete choice model that captures decision-making within affectively polarized social networks and propose a statistical inference method estimate key parameters -- in-group love and out-group hate -- from social media data. Through empirical validation from online discussions about the COVID-19 pandemic, we demonstrate that our approach accurately captures real-world polarization dynamics and explains the rapid emergence of a partisan gap in attitudes towards masking and lockdowns. This framework allows for tracking affective polarization across contentious issues has broad implications for fostering constructive online dialogues in digital spaces.

Georgios Chochlakis, Alexandros Potamianos, Kristina Lerman et al.

In-context Learning (ICL) has become the primary method for performing natural language tasks with Large Language Models (LLMs). The knowledge acquired during pre-training is crucial for this few-shot capability, providing the model with task priors. However, recent studies have shown that ICL predominantly relies on retrieving task priors rather than "learning" to perform tasks. This limitation is particularly evident in complex subjective domains such as emotion and morality, where priors significantly influence posterior predictions. In this work, we examine whether this is the result of the aggregation used in corresponding datasets, where trying to combine low-agreement, disparate annotations might lead to annotation artifacts that create detrimental noise in the prompt. Moreover, we evaluate the posterior bias towards certain annotators by grounding our study in appropriate, quantitative measures of LLM priors. Our results indicate that aggregation is a confounding factor in the modeling of subjective tasks, and advocate focusing on modeling individuals instead. However, aggregation does not explain the entire gap between ICL and the state of the art, meaning other factors in such tasks also account for the observed phenomena. Finally, by rigorously studying annotator-level labels, we find that it is possible for minority annotators to both better align with LLMs and have their perspectives further amplified.

Bohan Jiang, Dawei Li, Zhen Tan et al.

Measuring the relative impact of CTs is important for prioritizing responses and allocating resources effectively, especially during crises. However, assessing the actual impact of CTs on the public poses unique challenges. It requires not only the collection of CT-specific knowledge but also diverse information from social, psychological, and cultural dimensions. Recent advancements in large language models (LLMs) suggest their potential utility in this context, not only due to their extensive knowledge from large training corpora but also because they can be harnessed for complex reasoning. In this work, we develop datasets of popular CTs with human-annotated impacts. Borrowing insights from human impact assessment processes, we then design tailored strategies to leverage LLMs for performing human-like CT impact assessments. Through rigorous experiments, we textit{discover that an impact assessment mode using multi-step reasoning to analyze more CT-related evidence critically produces accurate results; and most LLMs demonstrate strong bias, such as assigning higher impacts to CTs presented earlier in the prompt, while generating less accurate impact assessments for emotionally charged and verbose CTs.

Kai Chen, Zihao He, Taiwei Shi et al.

Steerability, or the ability of large language models (LLMs) to adapt outputs to align with diverse community-specific norms, perspectives, and communication styles, is critical for real-world applications but remains under-evaluated. We introduce Steer-Bench, a benchmark for assessing population-specific steering using contrasting Reddit communities. Covering 30 contrasting subreddit pairs across 19 domains, Steer-Bench includes over 10,000 instruction-response pairs and validated 5,500 multiple-choice question with corresponding silver labels to test alignment with diverse community norms. Our evaluation of 13 popular LLMs using Steer-Bench reveals that while human experts achieve an accuracy of 81% with silver labels, the best-performing models reach only around 65% accuracy depending on the domain and configuration. Some models lag behind human-level alignment by over 15 percentage points, highlighting significant gaps in community-sensitive steerability. Steer-Bench is a benchmark to systematically assess how effectively LLMs understand community-specific instructions, their resilience to adversarial steering attempts, and their ability to accurately represent diverse cultural and ideological perspectives.

Woojae Jeong, Aditya Kommineni, Kleanthis Avramidis et al.

Depression and suicidality affect cognitive and emotional processes, yet objective, task-evoked neural readouts of mental health remain limited. We investigated the spatiotemporal dynamics of affective semantic processing using multivariate decoding of time-resolved, 64-channel electroencephalography (EEG). Participants (N=137) performed a sentence-evaluation task with emotionally salient, self-referential statements. We identified robust neural signatures of semantic processing, with peak decoding accuracy between 300-600 ms -- a window associated with rapid, stimulus-driven semantic evaluation and conflict monitoring. Relative to healthy controls, individuals with depression and suicidal ideation showed earlier onset, longer duration, and greater amplitude decoding responses, along with broader cross-temporal generalization and enhanced contributions from frontocentral and parietotemporal components. These findings suggest altered sensitivity and impaired disengagement from emotionally salient content in the clinical groups, advancing our understanding of the neurocognitive basis of mental health and establishing a compact and interpretable EEG-based index of semantic-evaluation dynamics with potential diagnostic relevance.

Siyi Guo, Myrl G. Marmarelis, Fred Morstatter et al.

Quantifying the effects of textual interventions in social systems, such as reducing anger in social media posts to see its impact on engagement, is challenging. Real-world interventions are often infeasible, necessitating reliance on observational data. Traditional causal inference methods, typically designed for binary or discrete treatments, are inadequate for handling the complex, high-dimensional textual data. This paper addresses these challenges by proposing CausalDANN, a novel approach to estimate causal effects using text transformations facilitated by large language models (LLMs). Unlike existing methods, our approach accommodates arbitrary textual interventions and leverages text-level classifiers with domain adaptation ability to produce robust effect estimates against domain shifts, even when only the control group is observed. This flexibility in handling various text interventions is a key advancement in causal estimation for textual data, offering opportunities to better understand human behaviors and develop effective interventions within social systems.

Minh Duc Chu, Kshitij Pawar, Zihao He et al.

Social media platforms increasingly struggle to detect harmful content that promotes muscle dysmorphic behaviors, particularly pro-bigorexia content that disproportionately affects adolescent males. Unlike traditional eating disorder detection focused on the "thin ideal," pro-bigorexia material masquerades as legitimate fitness content through complex multimodal combinations of visual displays, coded language, and motivational messaging that evade text-based detection systems. We address this challenge by developing BigTokDetect, a clinically-informed detection framework for identifying pro-bigorexia content on TikTok. We introduce BigTok, the first expert-annotated multimodal dataset of over 2,200 TikTok videos labeled by clinical psychologists and psychiatrists across five primary categories spanning body image, nutrition, exercise, supplements, and masculinity. Through a comprehensive evaluation of state-of-the-art vision language models, we achieve 82.9% accuracy on primary category classification and 69.0% on subcategory detection via domain-specific finetuning. Our ablation studies demonstrate that multimodal fusion improves performance by 5-10% over text-only approaches, with video features providing the most discriminative signals. These findings establish new benchmarks for multimodal harmful content detection and provide both the computational tools and methodological framework needed for scalable content moderation in specialized mental health domains.

Ashwin Rao, Sze Yuh Nina Wang, Kristina Lerman

The U.S. Supreme Court's 2022 ruling in Dobbs v. Jackson Women's Health Organization marked a turning point in the national debate over reproductive rights. While the ideological divide over abortion is well documented, less is known about how gender and local sociopolitical contexts interact to shape public discourse. Drawing on nearly 10 million abortion-related posts on X (formerly Twitter) from users with inferred gender, ideology and location, we show that gender significantly moderates abortion attitudes and emotional expression, particularly in conservative regions, and independently of ideology. This creates a gender gap in abortion attitudes that grows more pronounced in conservative regions. The leak of the Dobbs draft opinion further intensified online engagement, disproportionately mobilizing pro-abortion women in areas where access was under threat. These findings reveal that abortion discourse is not only ideologically polarized but also deeply structured by gender and place, highlighting the central role of identity in shaping political expression during moments of institutional disruption.

Kleanthis Avramidis, Woojae Jeong, Aditya Kommineni et al.

Identifying physiological and behavioral markers for mental health conditions is a longstanding challenge in psychiatry. Depression and suicidal ideation, in particular, lack objective biomarkers, with screening and diagnosis primarily relying on self-reports and clinical interviews. Here, we investigate eye tracking as a potential marker modality for screening purposes. Eye movements are directly modulated by neuronal networks and have been associated with attentional and mood-related patterns; however, their predictive value for depression and suicidality remains unclear. We recorded eye-tracking sequences from 126 young adults as they read and responded to affective sentences, and subsequently developed a deep learning framework to predict their clinical status. The proposed model included separate branches for trials of positive and negative sentiment, and used 2D time-series representations to account for both intra-trial and inter-trial variations. We were able to identify depression and suicidal ideation with an area under the receiver operating curve (AUC) of 0.793 (95% CI: 0.765-0.819) against healthy controls, and suicidality specifically with 0.826 AUC (95% CI: 0.797-0.852). The model also exhibited moderate, yet significant, accuracy in differentiating depressed from suicidal participants, with 0.609 AUC (95% CI 0.571-0.646). Discriminative patterns emerge more strongly when assessing the data relative to response generation than relative to the onset time of the final word of the sentences. The most pronounced effects were observed for negative-sentiment sentences, that are congruent to depressed and suicidal participants. Our findings highlight eye tracking as an objective tool for mental health assessment and underscore the modulatory impact of emotional stimuli on cognitive processes affecting oculomotor control.

Maxyn Leitner, Rebecca Dorn, Fred Morstatter et al.

The recent proliferation of short form video social media sites such as TikTok has been effectively utilized for increased visibility, communication, and community connection amongst trans/nonbinary creators online. However, these same platforms have also been exploited by right-wing actors targeting trans/nonbinary people, enabling such anti-trans actors to efficiently spread hate speech and propaganda. Given these divergent groups, what are the differences in network structure between anti-trans and pro-trans communities on TikTok, and to what extent do they amplify the effects of anti-trans content? In this paper, we collect a sample of TikTok videos containing pro and anti-trans content, and develop a taxonomy of trans related sentiment to enable the classification of content on TikTok, and ultimately analyze the reply network structures of pro-trans and anti-trans communities. In order to accomplish this, we worked with hired expert data annotators from the trans/nonbinary community in order to generate a sample of highly accurately labeled data. From this subset, we utilized a novel classification pipeline leveraging Retrieval-Augmented Generation (RAG) with annotated examples and taxonomy definitions to classify content into pro-trans, anti-trans, or neutral categories. We find that incorporating our taxonomy and its logics into our classification engine results in improved ability to differentiate trans related content, and that Results from network analysis indicate many interactions between posters of pro-trans and anti-trans content exist, further demonstrating targeting of trans individuals, and demonstrating the need for better content moderation tools

Zihao He, Minh Duc Chu, Rebecca Dorn et al.

Social scientists use surveys to probe the opinions and beliefs of populations, but these methods are slow, costly, and prone to biases. Recent advances in large language models (LLMs) enable the creating of computational representations or "digital twins" of populations that generate human-like responses mimicking the population's language, styles, and attitudes. We introduce Community-Cross-Instruct, an unsupervised framework for aligning LLMs to online communities to elicit their beliefs. Given a corpus of a community's online discussions, Community-Cross-Instruct automatically generates instruction-output pairs by an advanced LLM to (1) finetune a foundational LLM to faithfully represent that community, and (2) evaluate the alignment of the finetuned model to the community. We demonstrate the method's utility in accurately representing political and diet communities on Reddit. Unlike prior methods requiring human-authored instructions, Community-Cross-Instruct generates instructions in a fully unsupervised manner, enhancing scalability and generalization across domains. This work enables cost-effective and automated surveying of diverse online communities.

Ashwin Rao, Nazanin Sabri, Siyi Guo et al.

Effective communication during health crises is critical, with social media serving as a key platform for public health experts (PHEs) to engage with the public. However, it also amplifies pseudo-experts promoting contrarian views. Despite its importance, the role of emotional and moral language in PHEs' communication during COVID-19 remains under explored. This study examines how PHEs and pseudo-experts communicated on Twitter during the pandemic, focusing on emotional and moral language and their engagement with political elites. Analyzing tweets from 489 PHEs and 356 pseudo-experts from January 2020 to January 2021, alongside public responses, we identified key priorities and differences in messaging strategy. PHEs prioritize masking, healthcare, education, and vaccines, using positive emotional language like optimism. In contrast, pseudo-experts discuss therapeutics and lockdowns more frequently, employing negative emotions like pessimism and disgust. Negative emotional and moral language tends to drive engagement, but positive language from PHEs fosters positivity in public responses. PHEs exhibit liberal partisanship, expressing more positivity towards liberals and negativity towards conservative elites, while pseudo-experts show conservative partisanship. These findings shed light on the polarization of COVID-19 discourse and underscore the importance of strategic use of emotional and moral language by experts to mitigate polarization and enhance public trust.

Keith Burghardt, Kai Chen, Kristina Lerman

Adversarial information operations can destabilize societies by undermining fair elections, manipulating public opinions on policies, and promoting scams. Despite their widespread occurrence and potential impacts, our understanding of influence campaigns is limited by manual analysis of messages and subjective interpretation of their observable behavior. In this paper, we explore whether these limitations can be mitigated with large language models (LLMs), using GPT-3.5 as a case-study for coordinated campaign annotation. We first use GPT-3.5 to scrutinize 126 identified information operations spanning over a decade. We utilize a number of metrics to quantify the close (if imperfect) agreement between LLM and ground truth descriptions. We next extract coordinated campaigns from two large multilingual datasets from X (formerly Twitter) that respectively discuss the 2022 French election and 2023 Balikaran Philippine-U.S. military exercise in 2023. For each coordinated campaign, we use GPT-3.5 to analyze posts related to a specific concern and extract goals, tactics, and narrative frames, both before and after critical events (such as the date of an election). While the GPT-3.5 sometimes disagrees with subjective interpretation, its ability to summarize and interpret demonstrates LLMs' potential to extract higher-order indicators from text to provide a more complete picture of the information campaigns compared to previous methods.

Siyi Guo, Keith Burghardt, Valeria Pantè et al.

Social media user representation learning aims to capture user preferences, interests, and behaviors in low-dimensional vector representations. These representations are critical to a range of social problems, including predicting user behaviors and detecting inauthentic accounts. However, existing methods are either designed for commercial applications, or rely on specific features like text contents, activity patterns, or platform metadata, failing to holistically model user behavior across different modalities. To address these limitations, we propose SoMeR, a Social Media user Representation learning framework that incorporates temporal activities, text contents, profile information, and network interactions to learn comprehensive user portraits. SoMeR encodes user post streams as sequences of time-stamped textual features, uses transformers to embed this along with profile data, and jointly trains with link prediction and contrastive learning objectives to capture user similarity. We demonstrate SoMeR's versatility through three applications: 1) Identifying information operation driver accounts, 2) Measuring online polarization after major events, and 3) Predicting future user participation in Reddit hate communities. SoMeR provides new solutions to better understand user behavior in the socio-political domains, enabling more informed decisions and interventions.

Minh Duc Chu, Zihao He, Rebecca Dorn et al.

Eating disorders (ED), a severe mental health condition with high rates of mortality and morbidity, affect millions of people globally, especially adolescents. The proliferation of online communities that promote and normalize ED has been linked to this public health crisis. However, identifying harmful communities is challenging due to the use of coded language and other obfuscations. To address this challenge, we propose a novel framework to surface implicit attitudes of online communities by adapting large language models (LLMs) to the language of the community. We describe an alignment method and evaluate results along multiple dimensions of semantics and affect. We then use the community-aligned LLM to respond to psychometric questionnaires designed to identify ED in individuals. We demonstrate that LLMs can effectively adopt community-specific perspectives and reveal significant variations in eating disorder risks in different online communities. These findings highlight the utility of LLMs to reveal implicit attitudes and collective mindsets of communities, offering new tools for mitigating harmful content on social media.

Zihao He, Jonathan May, Kristina Lerman

Detecting norm violations in online communities is critical to maintaining healthy and safe spaces for online discussions. Existing machine learning approaches often struggle to adapt to the diverse rules and interpretations across different communities due to the inherent challenges of fine-tuning models for such context-specific tasks. In this paper, we introduce Context-aware Prompt-based Learning for Norm Violation Detection (CPL-NoViD), a novel method that employs prompt-based learning to detect norm violations across various types of rules. CPL-NoViD outperforms the baseline by incorporating context through natural language prompts and demonstrates improved performance across different rule types. Significantly, it not only excels in cross-rule-type and cross-community norm violation detection but also exhibits adaptability in few-shot learning scenarios. Most notably, it establishes a new state-of-the-art in norm violation detection, surpassing existing benchmarks. Our work highlights the potential of prompt-based learning for context-sensitive norm violation detection and paves the way for future research on more adaptable, context-aware models to better support online community moderators.

Julie Jiang, Kristina Lerman, Emilio Ferrara

While developments in machine learning led to impressive performance gains on big data, many human subjects data are, in actuality, small and sparsely labeled. Existing methods applied to such data often do not easily generalize to out-of-sample subjects. Instead, models must make predictions on test data that may be drawn from a different distribution, a problem known as \textit{zero-shot learning}. To address this challenge, we develop an end-to-end framework using a meta-learning approach, which enables the model to rapidly adapt to a new prediction task with limited training data for out-of-sample test data. We use three real-world small-scale human subjects datasets (two randomized control studies and one observational study), for which we predict treatment outcomes for held-out treatment groups. Our model learns the latent treatment effects of each intervention and, by design, can naturally handle multi-task predictions. We show that our model performs the best holistically for each held-out group and especially when the test group is distinctly different from the training group. Our model has implications for improved generalization of small-size human studies to the wider population.

Keith Burghardt, Kristina Lerman

Algorithms that aid human tasks, such as recommendation systems, are ubiquitous. They appear in everything from social media to streaming videos to online shopping. However, the feedback loop between people and algorithms is poorly understood and can amplify cognitive and social biases (algorithmic confounding), leading to unexpected outcomes. In this work, we explore algorithmic confounding in collaborative filtering-based recommendation algorithms through teacher-student learning simulations. Namely, a student collaborative filtering-based model, trained on simulated choices, is used by the recommendation algorithm to recommend items to agents. Agents might choose some of these items, according to an underlying teacher model, with new choices then fed back into the student model as new training data (approximating online machine learning). These simulations demonstrate how algorithmic confounding produces erroneous recommendations which in turn lead to instability, i.e., wide variations in an item's popularity between each simulation realization. We use the simulations to demonstrate a novel approach to training collaborative filtering models that can create more stable and accurate recommendations. Our methodology is general enough that it can be extended to other socio-technical systems in order to better quantify and improve the stability of algorithms. These results highlight the need to account for emergent behaviors from interactions between people and algorithms.

Yuzi He, Christopher Tran, Julie Jiang et al.

The popularity of online gaming has grown dramatically, driven in part by streaming and the billion-dollar e-sports industry. Online games regularly update their software to fix bugs, add functionality that improve the game's look and feel, and change the game mechanics to keep the games fun and challenging. An open question, however, is the impact of these changes on player performance and game balance, as well as how players adapt to these sudden changes. To address these questions, we use causal inference to measure the impact of software patches to League of Legends, a popular team-based multiplayer online game. We show that game patches have substantially different impacts on players depending on their skill level and whether they take breaks between games. We find that the gap between good and bad players increases after a patch, despite efforts to make gameplay more equal. Moreover, longer between-game breaks tend to improve player performance after patches. Overall, our results highlight the utility of causal inference, and specifically heterogeneous treatment effect estimation, as a tool to quantify the complex mechanisms of game balance and its interplay with players' performance.

Zihao He, Leili Tavabi, Kristina Lerman et al.

Dialogue Act (DA) classification is the task of classifying utterances with respect to the function they serve in a dialogue. Existing approaches to DA classification model utterances without incorporating the turn changes among speakers throughout the dialogue, therefore treating it no different than non-interactive written text. In this paper, we propose to integrate the turn changes in conversations among speakers when modeling DAs. Specifically, we learn conversation-invariant speaker turn embeddings to represent the speaker turns in a conversation; the learned speaker turn embeddings are then merged with the utterance embeddings for the downstream task of DA classification. With this simple yet effective mechanism, our model is able to capture the semantics from the dialogue content while accounting for different speaker turns in a conversation. Validation on three benchmark public datasets demonstrates superior performance of our model.