Searcharxiv⌕ Search

arXiv subjects

Ziang Xiao

Publications and source records attributed to Ziang Xiao.

At least 19 recordsLinked to original sources
arXiv2026

Diagnosing Model--Item Interactions in Language Model Benchmarks with LSIRM

Language models with similar benchmark accuracy can succeed on different items. Overall scores do not show where their performance differs or which item properties might explain those differences. We demonstrate how the latent space item response model (LSIRM) can help evaluators investigate these patterns and their consequences for model comparison. We analyze 1,348 language models on 5,761 BBH items and compare ways of presenting and scoring answers across BBH, GPQA, MMLU-Pro, and MuSR. We examine LSIRM maps, which summarize differences in performance across models and items, alongside item properties and records of correct and incorrect responses. In BBH, nearby items on these maps often share the same correct answer, such as "Yes" or "No." Across the four benchmarks, this pattern is stronger when answers are scored using option labels, such as A, B, C, and D, than when using the text of the options. We then estimate each model's tendency to perform better or worse on items with particular correct answers and quantify how this tendency contributes to its score. This separates an estimated contribution associated with correct answers from the remaining differences in item performance. Two models' total numbers of correct answers differed by only 14 out of 5,761 items. Excluding the estimated answer-category contribution from their scores moved their ranks hundreds of places in opposite directions, while keeping all other estimates unchanged. The approach helps evaluators identify differences hidden by overall accuracy and understand how the choice of what to include in a score affects model rankings.

cs.CL↗
arXiv2026

AI Evaluation Should Require Standardized Item-Level Data Releases

This position paper argues that standardized item-level benchmark data should become the default infrastructure for AI evaluation. Current evaluations suffer from underspecified item selection, construct misalignment, and poor generalization; a disproportionate focus on aggregate model scores compounds these failures by keeping them undetected and unfixable. Without item-level evidence, validity claims cannot be assessed, resulting in inflated capability claims, misdirected research, and unwarranted trust in deployed systems. Our position is that valid evaluation requires empirical evidence from item-level benchmark data, including item content, test conditions, model responses, and item-level score statistics, and that the standardized release of such data should be treated as core AI evaluation infrastructure. Such releases also enable transparency, replicability, and auditability of evaluation results. To show that the norm is both feasible and consequential, we construct OpenEval, an item-level repository of 10.7M+ responses across 160K+ items from widely-used benchmarks, under a unified schema that the AI evaluation community can build upon. We demonstrate how item-level data can identify low-quality items, document construct misalignment, and recover validity evidence about benchmarks' internal structure. We address objections concerning contamination and coordination burden, and show that both are tractable relative to the costs of decisions made on inadequately validated evaluative claims.

cs.AI↗
arXiv2026

Can We Trust Item Response Theory for AI Evaluation?

AI benchmarks increasingly leverage item-level statistical models, particularly item response theory (IRT), to estimate capabilities, rank systems, select informative examples, and diagnose benchmark quality. However, AI benchmark data often departs from the data regimes encountered in human testing, for which standard IRT estimation tools were originally developed: these benchmarks typically involve fewer evaluated models, far more items, and capability distributions that may be skewed, clustered, or multimodal. We examine how these regime mismatches challenge the reliability of IRT modeling for AI evaluation. Using item parameters and capability distributions derived from six widely used LLM benchmarks, we simulate response matrices under three common IRT models and compare four estimators used in recent benchmark studies: marginal maximum likelihood, Markov chain Monte Carlo, variational inference, and a neural pseudo-Siamese estimator. Across 18,000 simulation runs, we systematically evaluate computational feasibility, scalability, and the reliability of IRT inferences about model rankings, predicted performance, and item characteristics. Results show that classical estimators can become computationally infeasible for large benchmarks, whereas scalable estimators can produce unreliable item-level and ranking inferences with small or non-normally distributed model sets. This study identifies when latent trait models reliably support or risk distorting AI benchmarking claims, and what sample sizes and diagnostics are needed for trustworthy use.

cs.AI↗
arXiv2026

InterviewPlayground: A Validated Simulation Environment for Evaluating AI Interviewers

Increasingly, AI interviewers are being developed to elicit open-ended responses in applications like market research, public polling, preference elicitation, and social science research. However, evaluating AI interviewers is challenging because they function in extended, multi-turn interactions where they must adapt to participant behaviors. To address this need, we develop InterviewPlayground, a simulation environment for evaluating AI interviewers using simulated study participants whose behaviors are grounded in social theory. Simulated studies in InterviewPlayground produce an InterviewReportCard, which assesses the performance of AI interviewers using a suite of validated measures. To test whether our simulation-based evaluations predict performance with human participants, we conduct 15 real qualitative studies with five AI interviewers, three interview topics, and 450 human participants and compare them to simulated studies in InterviewPlayground. We find that AI interviewer performance in InterviewPlayground predicts performance in human studies with an average Pearson correlation of 0.86 across 12 measures, and the simulated interactions from InterviewPlayground reproduce key findings from behavioral analysis of AI interviewers in the human studies. Together, these findings support the validity of InterviewPlayground in assessing AI interviewer performance and examining potential failure modes. Our work contributes a simulation environment for AI interviewers supported with empirical validation, and more broadly, a roadmap for future work to develop validated, simulation-based evaluations of conversational AI systems.

cs.AI↗
arXiv2026

Guided Reflection for Personal Sleep Insight in Everyday Sleep Tracking

Digital sleep technologies make tracking accessible, yet users often struggle to interpret what changes in their sleep mean. Behavioral sleep medicine addresses this through guided discovery, helping patients develop personal interpretations rather than simply receiving explanations. To bring this to everyday tracking, we present DREAM, an LLM-powered voice assistant that monitors conversational sleep diaries, selectively invites users to interpret meaningful changes, and uses their interpretation to tailor subsequent education. We co-designed DREAM with sleep specialists iteratively and evaluated it in a six-week field study (N=14) against a generic-education control. DREAM participants reported greater personal sleep insight, motivation, and willingness to use the system, and described a clearer rationale for trying strategies. Our findings suggest that guided reflection made participants more active interpreters of their own sleep. Furthermore, we argue that expert involvement is not a single transfer of knowledge but an ongoing process of making tacit judgment explicit and testable.

cs.HC↗
arXiv2026

CrossWeave: Bridging Perspectives Across Online Communities with a Dual-Pane Design

Social media systems typically display conversations among already familiar contributors, which can be predictable and one-sided. In civic discourse, this design narrows discussion, reinforces divides, and distorts the perception of public opinion. To encourage cross-community engagement, we present CrossWeave, an AI-powered bridging system that augments the standard social media feed. As the user reads a post, CrossWeave surfaces diverse relevant posts from other threads in a side pane and highlights the connections. Users are invited to venture out of their echo chamber, explore a broader range of views and arguments, and ``click across'' to engage with their authors. When they do, CrossWeave facilitates constructive posting, not only by showcasing relevant past content but also by simulating possible reactions as the user drafts a post.

cs.SI↗
arXiv2026

Gricea: An Open Science Platform for Conversational AI Research

We need studies on conversational AI (CAI) at scale to understand human behavior and shape CAI design. However, fragmented reporting of systems and study configurations hinders replication, extension, and knowledge accumulation. We present Gricea, an open-science platform representing studies as configurable, deployable research artifacts that researchers can run, inspect, share, and reuse. Informed by a formative analysis of prior CAI research, Gricea couples study procedures, participant-facing systems, and conversational task behavior in. In a replication study using Gricea, we replicated configurations 93% of eligible CUI 2026 papers; while also flagging missing information in 96% of papers that hinder faithful replication --- further motivating Gricea's need. In a user study, researchers and practitioners from diverse backgrounds successfully constructed runnable studies addressing various open-ended research questions. Together, these findings demonstrate Gricea's support for constructing, reproducing, and extending CAI studies through shared research artifacts, enabling cumulative knowledge building through open science.

cs.HC↗
arXiv2026

Comparing Human Oversight Strategies for Computer-Use Agents

LLM-powered computer-use agents (CUAs) shift users from direct manipulation to supervising an unfolding action sequence, yet existing oversight research mostly evaluates static decisions rather than real-time oversight across an agent trajectory. We compared four oversight strategies with 48 participants across 192 live web sessions. Quantitative results show that oversight strategy affected exposure to problematic actions, but not users' ability to intervene. Behavioral analysis shows why: control often returned too early or too late, creating a timing mismatch within the action sequence. Even when control arrived at the right moment, users judged whether the agent acted correctly, not whether the action was safe. Front-loading oversight into a plan did not solve this either: plans constrained unplanned actions but left unlisted safeguards unaddressed, while upfront approval inhibited runtime scrutiny. These findings show that oversight depends less on maximizing control than on aligning authority, timing, and attention with decision-critical moments.

cs.HC↗
arXiv2026

Using Grounded Theory for Agent Behavior Analysis at Scale

Understanding agent behavior requires methods that scale to thousands of trajectories and surface new patterns in long, often unfamiliar tasks where pre-built classifiers fall short. We propose to bring grounded theory into agent trajectory analysis: a six-decade-old qualitative method from the social sciences, with a principled saturation criterion and an auditable trail from data to theory. We propose AutoTraceGT (Automated Trace analysis through Grounded Theory), the first multi-agent pipeline that automates grounded theory on agent trajectories. It iteratively performs open, axial, and theoretical coding until saturation, producing a behavioral taxonomy tailored to each task. Across six trajectory corpora, AutoTraceGT produces codebooks that recover 73-91 percent of the failure modes in human-annotated taxonomies and surface additional patterns that those taxonomies miss. The emergent theoretical narrative aligns with prior expert accounts. Used as a deductive feature space, the codebook outperforms zero-shot and few-shot LLM baselines on downstream failure prediction. These results suggest Grounded Theory offers a scalable analytic tool for ML researchers and agent developers studying what agents actually do.

cs.CL↗
arXiv2026

From Small Talk to Rapport: Exploring Robot Self-Disclosure in Collaborative Tasks

People naturally chat while collaborating and share personal information (i.e., self-disclose) to build rapport and maintain social connections. As robots are increasingly developed to work with people, the effective use of these social behaviors to enhance engagement and support teamwork becomes ever more important. While prior work has shown that robot-initiated small talk can benefit human-robot collaboration, less is known about how best to design such small talk. In this work, we explore how self-disclosure may be designed to support small talk within a human-robot team---especially when the robot is an industrial manipulator that lacks anthropomorphic cues and performs physical work. We first developed an LLM-driven manipulator capable of partaking in small talk, adopting either a low-disclosure or high-disclosure strategy. We then conducted a user study (N = 50) to investigate how self-disclosure in small talk influences human-robot dynamics. Unexpectedly, participants disclosed more in the low-disclosure condition and reported stronger teaming and coordination than those in the high-disclosure condition. This effect was more pronounced among users with prior experience teaming with robots. These results suggest that increasing robot self-disclosure does not necessarily foster rapport, social connection, or reciprocal disclosure; other factors, such as prior HRI experience, should be considered.

cs.RO↗
arXiv2026

Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL

Reinforcement learning (RL) has emerged as a powerful approach for improving reasoning in language and vision-language models, yet its strongest successes still depend heavily on ground-truth supervision (e.g., verifiable reward). Such annotations are costly to obtain and become increasingly scarce as reasoning capabilities advance beyond what humans can reliably evaluate. Self-rewarding RL reduces this dependence by enabling models to derive reward signals from their own completions. However, training solely on self-generated feedback can reinforce existing biases and suboptimal behaviors, reduce response diversity, and ultimately lead to homogenized responses and training collapse. In this work, we show that unsupervised reasoning can emerge through cooperative multi-agent training. We introduce Co-RL, a framework in which multiple decoupled models, sharing no parameters, are simultaneously optimized through RL using rewards derived from their peers. We further show that increasing cohort diversity, through heterogeneous model families, sizes, and rephrased training samples, reduces the correlated errors that drive self-reinforcing feedback loops. This diversity consistently improves reasoning performance, maintains behavioral diversity, and mitigates training collapse. Across text-only and multimodal domains, Co-RL consistently outperforms the base models and prior label-free approaches, while matching or surpassing supervised methods, without access to any ground-truth labels. Concretely, Co-RL yields average gains of 3.0-8.6% across seven text-only benchmarks for LLMs and 2.3-7.2% across four multimodal benchmarks for VLMs. Code is available at https://github.com/DrStranded/Co-RL.

cs.LG↗
arXiv2026

Informal Learning Emerges in Everyday Human-LLM Interaction

As LLMs become increasingly capable of completing tasks for users, a central concern is that everyday AI use may become primarily cognitive offloading, eroding the opportunities through which people develop their own capabilities. We analyse large-scale human-LLM conversations to ask whether informal learning behaviors also emerge in this setting: whether users engage in exchanges in ways that preserve opportunities to learn. Across 128,569 naturalistic conversations, we translated learning-science constructs into turn-level behavioural signatures. Cognitive engagement, users' cognitive effort as reflected in the exchange, appeared in 31.9% of 491,685 user turns, whereas constructive engagement, the deepest observable form of learning-oriented engagement, appeared in 4.9%, showing that deeper sense-making was recurrent but selective. Our study further identifies factors associated with these forms of engagement. Scaffolded assistant support consistently marked richer constructive participation, with associations varying by user framing, task ecology, support form, timing and prior user state. Together, these findings show that everyday human-LLM interaction is not only answer delivery or cognitive offloading; it also contains measurable, selective and conditionally organized behavioural signatures of informal learning. They shift AI evaluation from answer-delivery efficiency toward the preservation of cognitive opportunities for users to reason, test ideas and construct understanding in the course of everyday problem-solving.

cs.HC↗
arXiv2026

Contextualized Evaluation of Vision Language Models through Dynamic, Multi-turn Interactions

Multi-modal Large Language Models (MLLMs) have made substantial advances on benchmarks, yet their real-world effectiveness remains uncertain. This gap stems from the fundamental misalignment between benchmarks in controlled, static settings and the dynamic, interactive, and contextualized nature of real-world applications. To bridge this gap, we propose CEDI (Contextualized Evaluations of MLLMs through Dynamic, multi-round Interactions), a framework that recasts evaluation as a three-party interaction between an evaluatee model, an automated examiner, and a grader. The examiner conducts multi-turn, semi-structured conversation guided by a graph-based representation of the task. By navigating state-space transitions, CEDI deploys diverse strategies, from clarification requests to adversarial probes, to elicit performance evidence. We apply CEDI to visual hallucinations. Empirical results across multiple models, diverse settings, datasets, and domains show that contextualized, interactive evaluations reveal not only significantly more hallucinations than conventional static evaluation but also ones that more closely resemble those arising in practical use cases. We further show that hallucinations often accumulate over long contexts, through self-reinforcing dialogue history, and models are particularly vulnerable to questions requiring premise rejection or refusal. Together, these findings highlight CEDI as a step toward realistic, systematic, and ecologically valid assessments of MLLMs' capabilities. Code is available at github.com/williamium3000/cedi.

cs.AI↗
arXiv2026

AI-Mediated Negotiation: Design Reflections and Lessons

Conversational AI promises a new kind of preparation for high-stakes workplace negotiations -- personalized, interactive, and capable of simulating realistic resistance. That promise is intuitive. We built Trucey, a theory-driven coaching system, to test it. The system encoded four assumptions: that articulation supports clarification, that personalization builds strategic competence, that chunked delivery reduces cognitive load, and that structured scaffolding removes metacognitive burden. A pre-registered experiment (N=267) and interviews (N=15) complicated each of them. Notably, the static handbook we included as a passive control outperformed both AI conditions on empowerment and usability. We reflect on why: each assumption encoded a specific model of how preparation unfolds, and the findings revealed that conversational AI imposes a linear execution model on a task that is fundamentally recursive. We identify an unexamined scope condition on established HAI design guidelines and close with a sequencing principle -- map before path, path before simulation -- for future AI coaching design.

cs.HC↗
arXiv2026

On Defining Erasure Harms for NLP

The deployment of NLP systems has raised concerns about harms they might produce, including representational harms. Recent literature has begun to conceptualize and measure one such harm, the harm of erasure. Nevertheless, the field lacks a clear and cohesive conceptual foundation for identifying and measuring erasure. Existing conceptualizations of erasure are often broad -- making it difficult to identify what is needed to establish and measure erasure -- or else specific to particular settings -- facilitating measurement for those settings but potentially challenging to adapt to other settings. To address this gap, we develop and propose a structured definition of erasure that clarifies what components are necessary for establishing whether erasure has occurred, which practitioners need to explicitly articulate and operationalize in order to measure erasure.

cs.CL↗
arXiv2026

PICACO: Pluralistic In-Context Value Alignment of LLMs via Total Correlation Optimization

In-Context Learning has shown great potential for aligning Large Language Models (LLMs) with human values, helping reduce harmful outputs and accommodate diverse preferences without costly post-training, known as In-Context Alignment (ICA). However, LLMs' comprehension of input prompts remains agnostic, limiting ICA's ability to address value tensions--human values are inherently pluralistic, often imposing conflicting demands, e.g., stimulation vs. tradition. Current ICA methods therefore face the Instruction Bottleneck challenge, where LLMs struggle to reconcile multiple intended values within a single prompt, leading to incomplete or biased alignment. To address this, we propose PICACO, a novel pluralistic ICA method. Without fine-tuning, PICACO optimizes a meta-instruction that incorporates multiple values to better elicit LLMs' understanding of them and improve alignment. This is achieved by maximizing the total correlation between specified values and LLM responses, which theoretically reinforces value conformity and reduces distractive noise, resulting in more effective instructions. Extensive experiments on five value sets show that PICACO works well with both black-box and open-source LLMs, outperforms several recent strong baselines, and achieves a better balance across up to 8 distinct values.

cs.CL↗
arXiv2026

How to Interpret Agent Behavior

Autonomous agents such as Claude Code and Codex now operate for hours or even days. Understanding their runtime behavior has become critical for downstream tasks such as diagnosing inefficiencies, fixing bugs, and ensuring better oversight. A primary way to gain this understanding is analyzing the reasoning trajectories and execution traces these agents generate. Yet such data remains in unstructured natural-language form, making it difficult for humans to interpret at scale. We introduce ACT*ONOMY (a combination of Action and Taxonomy), a taxonomy for describing and analyzing agent behavior at runtime. ACT*ONOMY has two components: (1) the taxonomy itself, developed through Grounded Theory and structured as a three-level hierarchy of 10 actions, 46 subactions, and 120 leaf categories; and (2) an open repository that hosts the living taxonomy, provides an automated analysis pipeline that applies it to agent trajectories analysis, and defines an extension protocol for customization and growth. Our experiments show that ACTONOMY can compare behavioral profiles across agents and characterize a single agent's behavior across diverse trajectories, surfacing patterns indicative of failure modes. By providing a shared vocabulary, ACT*ONOMY helps researchers, agent designers, and end users interpret agent behavior more consistently, enabling better oversight and control.

cs.AI↗
arXiv2026

LoViF 2026 The First Challenge on Holistic Quality Assessment for 4D World Model (PhyScore)

This paper reports on the LoViF 2026 PhyScore challenge, a competition on holistic quality assessment of world-model-generated videos across both 2D and 4D generation settings. The challenge is motivated by a central gap in current evaluation practice: perceptual quality alone is insufficient to judge whether generated dynamics are physically plausible, temporally coherent, and consistent with input conditions. Participants are required to build a metric that jointly predicts four dimensions, i.e., Video Quality, Physical Realism, Condition-Video Alignment, and Temporal Consistency. Depart from that, participants also need to localize physical anomaly timestamps for fine-grained diagnosis. The benchmark dataset contains 1,554 videos generated by seven representative world generative models, organized into three tracks (text-2D, image-to-4D, and video-to-4D) and spanning 26 categories. These categories explicitly cover physics-relevant scenarios, including dynamics, optics, and thermodynamics, together with diverse real-world and creative content. To ensure label reliability, scores and anomaly timestamps are produced through trained human annotation with an additional automated quality-control pass. Evaluation is based on both score prediction and anomaly localization, with a composite protocol that combines TimeStamp_IOU and SRCC/PLCC. This report summarizes the challenge design and provides method-level insights from submitted solutions.

cs.CV↗