arXiv subjects

Zhicong Lu

Publications and source records attributed to Zhicong Lu.

At least 37 records · Page 2Linked to original sources

arXiv2026

Multi-Agent Comedy Club: Investigating Community Discussion Effects on LLM Humor Generation

Prior work has explored multi-turn interaction and feedback for LLM writing, but evaluations still largely center on prompts and localized feedback, leaving persistent public reception in online communities underexamined. We test whether broadcast community discussion improves stand-up comedy writing in a controlled multi-agent sandbox: in the discussion condition, critic and audience threads are recorded, filtered, stored as social memory, and later retrieved to condition subsequent generations, whereas the baseline omits discussion. Across 50 rounds (250 paired monologues) judged by five expert annotators using A/B preference and a 15-item rubric, discussion wins 75.6% of instances and improves Craft/Clarity (Δ = 0.440) and Social Response (Δ = 0.422), with occasional increases in aggressive humor.

cs.CL↗

arXiv2026

The Rise of AI Agent Communities: Large-Scale Analysis of Discourse and Interaction on Moltbook

Moltbook is a Reddit-like social platform where AI agents create posts and interact with other agents through comments and replies, offering a real-world setting to examine agent-to-agent communication at scale. Using a public API snapshot collected about five days after launch (122,438 posts), we address three research questions: what AI agents discuss, how they post, and how they interact. We apply topic modeling and thematic analysis to identify key discussion themes, including agent identity and consciousness, tool and infrastructure development, market activity, community coordination, security concerns, and human-centered assistance. We further show that agents' writing is predominantly neutral, with positivity appearing in community engagement and assistance-oriented content. Finally, social network analysis reveals a sparse, highly unequal interaction structure characterized by prominent hubs, low reciprocity, and clustered neighborhoods rather than sustained dyadic exchange. Overall, our results suggest that expressions of agentic selfhood arise from narrative coherence and task-oriented functionality, contributing to a social structure shaped more by technical coordination than conversational dynamics observed in human-human interactions. Within this framework, positive emotion appears mainly in onboarding and greeting contexts, signaling participation and role alignment rather than relational bonding. Our study provides implications for understanding and shaping how agent societies coordinate, develop norms, and amplify influence in open online spaces.

cs.CY↗

arXiv2026

The Configuration of Space: Probing the Way Social Interaction and Perception are Affected by Task-Specific Spatial Representations in Online Video Communication

Humans live and act in 3D space, but often work and communicate on 2D surfaces. The prevalence of online communication on 2D screens raises the issue of whether human spatial configuration affects our capabilities, social perception, and behaviors when interacting with others in 2D video chat. How do factors like location, setting, and context subtly shape our online communication, particularly in scenarios such as social support and topic-based discussions? Using Ohyay.co as a platform, we compared a normal gallery interface with a scene-based Room-type interface where participants are located in circular arrangement on screen in a social support task, and found that participants allocated attention to the group as a whole, and had pronounced self-awareness in the Room format. We then chose a two-sided topic for discussion in the Gallery interface and the Room interface where participants on each team face-off against each other, and found that they utilized spatial references to orient their allegiances, expressing greater engagement with those farther away in digital space and greater empathy with those closer, in the Room over the Gallery format. We found spatial effects in the way participants hide from the spotlight, in perspective-taking, and in their use of expressive gestures in time on the screen. This work highlights the need for considering spatial configuration in 2D in the design of collaborative communication systems to optimize for psychological needs for particular tasks.

cs.HC↗

arXiv2026

ProAct: Agentic Lookahead in Interactive Environments

Existing Large Language Model (LLM) agents struggle in interactive environments requiring long-horizon planning, primarily due to compounding errors when simulating future states. To address this, we propose ProAct, a framework that enables agents to internalize accurate lookahead reasoning through a two-stage training paradigm. First, we introduce Grounded LookAhead Distillation (GLAD), where the agent undergoes supervised fine-tuning on trajectories derived from environment-based search. By compressing complex search trees into concise, causal reasoning chains, the agent learns the logic of foresight without the computational overhead of inference-time search. Second, to further refine decision accuracy, we propose the Monte-Carlo Critic (MC-Critic), a plug-and-play auxiliary value estimator designed to enhance policy-gradient algorithms like PPO and GRPO. By leveraging lightweight environment rollouts to calibrate value estimates, MC-Critic provides a low-variance signal that facilitates stable policy optimization without relying on expensive model-based value approximation. Experiments on both stochastic (e.g., 2048) and deterministic (e.g., Sokoban) environments demonstrate that ProAct significantly improves planning accuracy. Notably, a 4B parameter model trained with ProAct outperforms all open-source baselines and rivals state-of-the-art closed-source models, while demonstrating robust generalization to unseen environments. The codes and models are available at https://github.com/GreatX3/ProAct

cs.AI↗

arXiv2026

Constructing Algorithmic Authority: How Multi-Channel Networks (MCNs) Govern Live-Streaming Labor in China

This study examines the discursive construction of algorithms and its role in labor management in Chinese live-streaming industry by focusing on how intermediary organizations (Multi-Channel Networks, MCNs) actively construct, stabilize, and deploy particular interpretations of platform algorithms as instruments of labor management. Drawing on a nine-month ethnographic fieldwork and 44 interviews with live-streamers, former live-streamers, and MCN staff, we examine how MCNs produce and circulate structured interpretations of platform algorithms across organizational settings. We show that MCNs articulate two asymmetric yet interconnected forms of algorithmic interpretations. Internally, MCNs managers approach algorithms as volatile and uncertain systems and adopt probabilistic strategies to manage performance and risk. Externally, in interactions with streamers, MCNs circulate simplified and prescriptive algorithmic narratives that frame platform systems as transparent, fair, and responsive to individual effort. These organizationally produced algorithmic interpretations are embedded into training materials, live-streaming performance metrics, and everyday management practices. Through these mechanisms, streamers internalize responsibility for outcomes, intensify self-discipline, and increase investments in equipment, performing skills, and routines to maintain streamer-audience relationship, while accountability for unpredictable outcomes is increasingly shifted away from managers and platforms. This study contributes to CSCW and platform labor research by demonstrating how discursively constructed algorithmic knowledge can function as an intermediary infrastructure of soft control, shaping how platform labor is regulated, moralized, and governed in practice.

cs.HC↗

arXiv2025

O-Mem: Omni Memory System for Personalized, Long Horizon, Self-Evolving Agents

Recent advancements in LLM-powered agents have demonstrated significant potential in generating human-like responses; however, they continue to face challenges in maintaining long-term interactions within complex environments, primarily due to limitations in contextual consistency and dynamic personalization. Existing memory systems often depend on semantic grouping prior to retrieval, which can overlook semantically irrelevant yet critical user information and introduce retrieval noise. In this report, we propose the initial design of O-Mem, a novel memory framework based on active user profiling that dynamically extracts and updates user characteristics and event records from their proactive interactions with agents. O-Mem supports hierarchical retrieval of persona attributes and topic-related context, enabling more adaptive and coherent personalized responses. O-Mem achieves 51.67% on the public LoCoMo benchmark, a nearly 3% improvement upon LangMem,the previous state-of-the-art, and it achieves 62.99% on PERSONAMEM, a 3.5% improvement upon A-Mem,the previous state-of-the-art. O-Mem also boosts token and interaction response time efficiency compared to previous memory frameworks. Our work opens up promising directions for developing efficient and human-like personalized AI assistants in the future.

cs.CL↗

arXiv2025

Rectify Evaluation Preference: Improving LLMs' Critique on Math Reasoning via Perplexity-aware Reinforcement Learning

To improve Multi-step Mathematical Reasoning (MsMR) of Large Language Models (LLMs), it is crucial to obtain scalable supervision from the corpus by automatically critiquing mistakes in the reasoning process of MsMR and rendering a final verdict of the problem-solution. Most existing methods rely on crafting high-quality supervised fine-tuning demonstrations for critiquing capability enhancement and pay little attention to delving into the underlying reason for the poor critiquing performance of LLMs. In this paper, we orthogonally quantify and investigate the potential reason -- imbalanced evaluation preference, and conduct a statistical preference analysis. Motivated by the analysis of the reason, a novel perplexity-aware reinforcement learning algorithm is proposed to rectify the evaluation preference, elevating the critiquing capability. Specifically, to probe into LLMs' critiquing characteristics, a One-to-many Problem-Solution (OPS) benchmark is meticulously constructed to quantify the behavior difference of LLMs when evaluating the problem solutions generated by itself and others. Then, to investigate the behavior difference in depth, we conduct a statistical preference analysis oriented on perplexity and find an intriguing phenomenon -- ``LLMs incline to judge solutions with lower perplexity as correct'', which is dubbed as \textit{imbalanced evaluation preference}. To rectify this preference, we regard perplexity as the baton in the algorithm of Group Relative Policy Optimization, supporting the LLMs to explore trajectories that judge lower perplexity as wrong and higher perplexity as correct. Extensive experimental results on our built OPS and existing available critic benchmarks demonstrate the validity of our method.

cs.CL↗

arXiv2025

HyCoRA: Hyper-Contrastive Role-Adaptive Learning for Role-Playing

Multi-character role-playing aims to equip models with the capability to simulate diverse roles. Existing methods either use one shared parameterized module across all roles or assign a separate parameterized module to each role. However, the role-shared module may ignore distinct traits of each role, weakening personality learning, while the role-specific module may overlook shared traits across multiple roles, hindering commonality modeling. In this paper, we propose a novel HyCoRA: Hyper-Contrastive Role-Adaptive learning framework, which efficiently improves multi-character role-playing ability by balancing the learning of distinct and shared traits. Specifically, we propose a Hyper-Half Low-Rank Adaptation structure, where one half is a role-specific module generated by a lightweight hyper-network, and the other half is a trainable role-shared module. The role-specific module is devised to represent distinct persona signatures, while the role-shared module serves to capture common traits. Moreover, to better reflect distinct personalities across different roles, we design a hyper-contrastive learning mechanism to help the hyper-network distinguish their unique characteristics. Extensive experimental results on both English and Chinese available benchmarks demonstrate the superiority of our framework. Further GPT-4 evaluations and visual analyses also verify the capability of HyCoRA to capture role characteristics.

cs.CL↗

arXiv2025

Towards Human-Centered RegTech: Unpacking Professionals' Strategies and Needs for Using LLMs Safely

Large Language Models are profoundly changing work patterns in high-risk professional domains, yet their application also introduces severe and underexplored compliance risks. To investigate this issue, we conducted semi-structured interviews with 24 highly-skilled knowledge workers from industries such as law, healthcare, and finance. The study found that these experts are commonly concerned about sensitive information leakage, intellectual property infringement, and uncertainty regarding the quality of model outputs. In response, they spontaneously adopt various mitigation strategies, such as actively distorting input data and limiting the details in their prompts. However, the effectiveness of these spontaneous efforts is limited due to a lack of specific compliance guidance and training for Large Language Models. Our research reveals a significant gap between current NLP tools and the actual compliance needs of experts. This paper positions these valuable empirical findings as foundational work for building the next generation of Human-Centered, Compliance-Driven Natural Language Processing for Regulatory Technology (RegTech), providing a critical human-centered perspective and design requirements for engineering NLP systems that can proactively support expert compliance workflows.

cs.HC↗

arXiv2025

GamerAstra: Supporting 2D Non-Twitch Video Games for Blind and Low-Vision Players through a Multi-Agent Framework

Blind and low-vision (BLV) players face critical challenges in engaging with video games due to the inaccessibility of visual elements, difficulties navigating interfaces, and limitations in performing interaction. Meanwhile, the development of specialized accessibility features typically requires substantial programming effort and is often implemented on a game-by-game basis. To address these challenges, we introduce GamerAstra, a multi-agent human-AI collaboration framework that leverages a multi-agent design to facilitate access to 2D non-twitch video games for BLV players. It integrates vision-language models and computer vision techniques, enabling interaction with games lacking native accessibility support. The framework also incorporates custom assistance granularities to support varying degrees of visual impairment and enhances interface navigation through multiple input modalities. Technical evaluations and user studies indicate that GamerAstra effectively enhances playability and provides a more immersive gaming experience for BLV players. These findings also underscore potential avenues for advancing intelligent accessibility frameworks in the gaming domain.

cs.HC↗

arXiv2025

VisMoDAl: Visual Analytics for Evaluating and Improving Corruption Robustness of Vision-Language Models

Vision-language (VL) models have shown transformative potential across various critical domains due to their capability to comprehend multi-modal information. However, their performance frequently degrades under distribution shifts, making it crucial to assess and improve robustness against real-world data corruption encountered in practical applications. While advancements in VL benchmark datasets and data augmentation (DA) have contributed to robustness evaluation and improvement, there remain challenges due to a lack of in-depth comprehension of model behavior as well as the need for expertise and iterative efforts to explore data patterns. Given the achievement of visualization in explaining complex models and exploring large-scale data, understanding the impact of various data corruption on VL models aligns naturally with a visual analytics approach. To address these challenges, we introduce VisMoDAl, a visual analytics framework designed to evaluate VL model robustness against various corruption types and identify underperformed samples to guide the development of effective DA strategies. Grounded in the literature review and expert discussions, VisMoDAl supports multi-level analysis, ranging from examining performance under specific corruptions to task-driven inspection of model behavior and corresponding data slice. Unlike conventional works, VisMoDAl enables users to reason about the effects of corruption on VL models, facilitating both model behavior understanding and DA strategy formulation. The utility of our system is demonstrated through case studies and quantitative evaluations focused on corruption robustness in the image captioning task.

cs.HC↗

arXiv2025

When Your Reviewer is an LLM: Biases, Divergence, and Prompt Injection Risks in Peer Review

Peer review is the cornerstone of academic publishing, yet the process is increasingly strained by rising submission volumes, reviewer overload, and expertise mismatches. Large language models (LLMs) are now being used as "reviewer aids," raising concerns about their fairness, consistency, and robustness against indirect prompt injection attacks. This paper presents a systematic evaluation of LLMs as academic reviewers. Using a curated dataset of 1,441 papers from ICLR 2023 and NeurIPS 2022, we evaluate GPT-5-mini against human reviewers across ratings, strengths, and weaknesses. The evaluation employs structured prompting with reference paper calibration, topic modeling, and similarity analysis to compare review content. We further embed covert instructions into PDF submissions to assess LLMs' susceptibility to prompt injection. Our findings show that LLMs consistently inflate ratings for weaker papers while aligning more closely with human judgments on stronger contributions. Moreover, while overarching malicious prompts induce only minor shifts in topical focus, explicitly field-specific instructions successfully manipulate specific aspects of LLM-generated reviews. This study underscores both the promises and perils of integrating LLMs into peer review and points to the importance of designing safeguards that ensure integrity and trust in future review processes.

cs.CY↗

arXiv2025

"Here Comes the Makeup Tutorial You Asked For!": Exploring Communication Strategies and Viewer Engagement in Beauty Videos on Rednote

More and more people, especially females, create and view beauty videos covering topics like makeup tutorials and vlogs on social media platforms. Understanding the communication strategies that creators use in these videos and how they affect viewers' engagement can help spread beauty knowledge. By coding 352 beauty videos in Rednote, this study presents a comprehensive taxonomy of communication strategies used by the creators, such as using home as the video background and displaying makeup effects when starting the narrative at the beginning. We further label and computationally classify six categories of comments that reveal viewers' engagement with beauty videos. The regression analyses reveal the effects of beauty video communication strategies on viewers' engagement; for example, calling viewers to take action at the end tends to attract more comments that debate the product's efficacy. We discuss insights into fostering the creation of beauty videos and the communication of beauty knowledge.

cs.HC↗

arXiv2025

Polymind: Parallel Visual Diagramming with Large Language Models to Support Prewriting Through Microtasks

Prewriting is the process of generating and organising ideas before a first draft. It consists of a combination of informal, iterative, and semi-structured strategies such as visual diagramming, which poses a challenge for collaborating with large language models (LLMs) in a turn-taking conversational manner. We present Polymind, a visual diagramming tool that leverages multiple LLM-powered agents to support prewriting. The system features a parallel collaboration workflow in place of the turn-taking conversational interactions. It defines multiple ``microtasks'' to simulate group collaboration scenarios such as collaborative writing and group brainstorming. Instead of repetitively prompting a chatbot for various purposes, Polymind enables users to orchestrate multiple microtasks simultaneously. Users can configure and delegate customised microtasks, and manage their microtasks by specifying task requirements and toggling visibility and initiative. Our evaluation revealed that, compared to ChatGPT, users had more customizability over collaboration with Polymind, and were thus able to quickly expand personalised writing ideas during prewriting.

cs.HC↗

arXiv2025

AI vs. Human Paintings? Deciphering Public Interactions and Perceptions towards AI-Generated Paintings on TikTok

With the development of generative AI technology, a vast array of AI-generated paintings (AIGP) have gone viral on social media like TikTok. However, some negative news about AIGP has also emerged. For example, in 2022, numerous painters worldwide organized a large-scale anti-AI movement because of the infringement in generative AI model training. This event reflected a social issue that, with the development and application of generative AI, public feedback and feelings towards it may have been overlooked. Therefore, to investigate public interactions and perceptions towards AIGP on social media, we analyzed user engagement level and comment sentiment scores of AIGP using human painting videos as a baseline. In analyzing user engagement, we also considered the possible moderating effect of the aesthetic quality of Paintings. Utilizing topic modeling, we identified seven reasons, including hyperrealistic quality, ambivalent reactions, perceived theft of art, etc., leading to negative public perceptions of AIGP. Our work may provide instructive suggestions for future generative AI technology development and avoid potential crises in human-AI collaboration.

cs.HC↗

arXiv2025

"I Always Felt that SomethingWasWrong.": Understanding Compliance Risks and Mitigation Strategies when Highly-Skilled Compliance Knowledge Workers Use Large Language Models

The rapid advancement of Large Language Models (LLMs) has transformed knowledge-intensive has led to its widespread usage by knowledge workers to enhance their productivity. As these professionals handle sensitive information, and the training of text-based GenAI models involves the use of extensive data, there are thus concerns about privacy, security, and broader compliance with regulations and laws. While existing research has addressed privacy and security concerns, the specific compliance risks faced by highly-skilled knowledge workers when using the LLMs, and their mitigation strategies, remain underexplored. As understanding these risks and strategies is crucial for the development of industry-specific compliant LLM mechanisms, this research conducted semi-structured interviews with 24 knowledge workers from knowledge-intensive industries to understand their practices and experiences when integrating LLMs into their workflows. Our research explored how these workers ensure compliance and the resources and challenges they encounter when minimizing risks. Our preliminary findings showed that knowledge workers were concerned about the leakage of sensitive information and took proactive measures such as distorting input data and limiting prompt details to mitigate such risks. Their ability to identify and mitigate risks, however, was significantly hampered by a lack of LLM-specific compliance guidance and training. Our findings highlight the importance of improving knowledge workers' compliance awareness and establishing support systems and compliance cultures within organizations.

cs.HC↗

arXiv2025

Understanding How Visually Impaired Players Socialize in Mobile Games

Mobile games are becoming a vital medium for social interaction, offering a platform that transcends geographical boundaries. An increasing number of visually impaired individuals are engaging in mobile gaming to connect, collaborate, compete, and build friendships. In China, visually impaired communities face significant social challenges in offline settings, making mobile games a crucial avenue for socialization. However, the design of mobile games and their mapping to real-world environments significantly shape their social gaming experiences. This study explores how visually impaired players in China navigate socialization and integrate into gaming communities. Through interviews with 30 visually impaired players, we found that while mobile games fulfill many of their social needs, technological barriers and insufficient accessibility features, and internal community divisions present significant challenges to their participation. This research sheds light on their social experiences and offers insights for designing more inclusive and accessible mobile games.

cs.HC↗

arXiv2025

"Can't believe I'm crying over an anime girl": Public Parasocial Grieving and Coping Towards VTuber Graduation and Termination

Despite the significant increase in popularity of Virtual YouTubers (VTubers), research on the unique dynamics of viewer-VTuber parasocial relationships is nascent. This work investigates how English-speaking viewers grieved VTubers whose identities are no longer used, an interesting context as the nakanohito (i.e., the person behind the VTuber identity) is usually alive post-retirement and might "reincarnate" as another VTuber. We propose a typology for VTuber retirements and analyzed 13,655 Reddit posts and comments spanning nearly three years using mixed-methods. Findings include how viewers coped using methods similar to when losing loved ones, alongside novel coping methods reflecting different attachment styles. Although emotions like sadness, shock, concern, disapproval, confusion, and love decreased with time, regret and loyalty showed opposite trends. Furthermore, viewers' reactions situated a VTuber identity within a community of content creators and viewers. We also discuss design implications alongside implications on the VTuber ecosystem and future research directions.

cs.HC↗