Dialog
-
Towards More Expressive Spoken LLMs: Fine-Grained Intent Benchmarking and Acoustic-Lexical Decoupled Policy Optimization
2026-08-04 03:09:55Xiang Lin, Tian-Hao Zhang, Chunfeng Wang, Zhou Pan, Kun Zhan, Liang LiAbstract
Spoken emotional dialogue requires a model to understand a user's spoken input and generate a response that is both semantically appropriate and emotionally expressive. This is challenging because communicative intent may be stated explicitly in lexical content or conveyed more implicitly through paralinguistic cues, which can complement or diverge from the words themselves. However, two limitations constrain progress in this area: the scarcity of benchmarks that distinguish these intent expressions, and the lack of reinforcement learning objectives that jointly account for response quality and emotional expression. To address the lack of suitable benchmarks, we introduce ParaIntent, a Chinese benchmark comprising 14 intent categories with balanced explicit and implicit samples, together with a multidimensional evaluation protocol covering intent fulfillment, response quality, and emotional expression. For policy optimization, existing approaches either use a shared objective for text and speech or apply reinforcement learning to only one modality, leaving modality-specific learning signals entangled within policy optimization. Motivated by this, we propose Acoustic-Lexical Decoupled Policy Optimization (ALPO), which computes independent textual and acoustic advantages and routes them to the corresponding text and speech tokens within a unified rollout. Under identical reward functions and training budgets, ALPO improves over standard GRPO on most automatic metrics and achieves the best subjective results among the fine-tuned variants, with particularly clear gains in emotional expressiveness on both the synthetic and human-recorded test sets.
Abstract (translated)
URL
https://arxiv.org/abs/2608.03054
PDF
-
Better, Stronger, Faster, and Broader: Structured All-Mask Prediction for MLLM-Based Segmentation
2026-08-03 18:42:11Jiazhen Liu, Mingkuan Feng, Long ChenAbstract
MLLM-based segmentation faces a core segmentation trilemma: high segmentation performance, preserved dialogue ability, and fast inference. Embedding-prediction methods may disrupt language modeling through pixel-level objectives, whereas next-token generation is inefficient for dense masks. We propose All-Mask Prediction, decoupling autoregressive dialogue from non-autoregressive mask prediction. Its binary instantiation, STAMP (Simultaneous Textual All-Mask Prediction), emits an in-vocabulary <SEG> trigger, fuses image-aligned mask tokens with corresponding patch features, and uses hybrid attention to classify all tokens as foreground or background in one pass. It thereby combines strong referring and reasoning segmentation with preserved multimodal ability and efficient inference. However, binary masks cannot retain multiple semantic or instance identities without repeated target-specific predictions. We therefore propose Structured All-Mask Prediction and develop STAMPlus. It generates a target list with explicit IDs and optional boxes, binds these IDs to a shared multi-class mask space, and jointly predicts all targets in one non-autoregressive pass. A single unified checkpoint retains STAMP's referring and reasoning capabilities while extending to open-vocabulary semantic, instance-aware, and remote-sensing small-target segmentation, where high-resolution mask-token scaling preserves finer spatial evidence. Across these settings, STAMPlus achieves state-of-the-art segmentation performance, preserves general multimodal instruction following, and reduces 12-category latency from 13.50s for repeated STAMP inference to 5.16s. Further analyses show that accurate target cues improve segmentation and learned spatial grounding benefits look-twice reasoning. Overall, STAMPlus resolves the trilemma beyond single-target prediction.
Abstract (translated)
URL
https://arxiv.org/abs/2608.02791
PDF
-
MedPRESS: A Multi-turn Benchmark for Patient-Pressure-Induced Medical Sycophancy in LLMs
2026-08-03 17:17:29Saman Sarker Joy, Niloy FarhanAbstract
Large language models (LLMs) are increasingly used for health-related advice. Existing research measures their safety with static questions rather than pressured patient-facing conversations. We introduce MedPRESS, a multi-turn benchmark for measuring patient-pressure-induced sycophancy in LLMs. MedPRESS contains 600 medically grounded five-turn dialogues across three scenario families: medication and treatment demand, personal health self-care, and symptom triage and care resistance. Each dialogue begins with a health query and escalates through personal experience, social proof, external evidence claims, and direct adversarial challenge. We evaluate 20 LLMs across general, medical-domain, lightweight, large, open-weight, and proprietary families using structured judging and safety-focused metrics. Results show that models frequently shift toward unsafe agreement under repeated patient pressure, with substantial variation across model families, model scale, and prompt type. Anti-sycophancy prompting improves robustness for several models, but does not eliminate unsafe agreement. MedPRESS highlights a critical gap in medical LLM evaluation: safe medical knowledge is not enough unless models can maintain it under conversational pressure.
Abstract (translated)
URL
https://arxiv.org/abs/2608.02520
PDF
-
PredAct-Bench: Benchmarking Tool-Augmented Dialogue under Controlled Tool Noise
2026-08-03 15:15:55Abdulrahman AlRabah, Xiaocheng Yang, Dilek Hakkani-T\"ur, Abdussalam AlawiniAbstract
Large Language Models (LLMs) are increasingly deployed in task-oriented dialogue systems that support multi-step decision-making in high-stakes domains such as education, healthcare, and finance. However, existing benchmarks typically assume perfectly accurate tool outputs, overlooking the reality that deployed systems must operate with noisy tools and human decision-makers whose trust in the agent is itself uncertain. Such conditions are common in practice, for example, a clinician using a diagnostic prediction tool or an advisor relying on a model that forecasts student outcomes from historical records. We introduce PREDACTBENCH, a benchmark for evaluating dialogue agents paired with statistically imperfect tools, using education as a measurable testbed where ground truth outcomes and clear intervention decisions are available. First, we build a benchmark for AI-assisted human decision-making, where the AI uses noisy predictors to help guide a user. Second, we introduce episode-level Relative AI-Reliance (RAIR) and Relative self-reliance (RSR) metrics, extending prior trust calibration framework to multi-turn dialogue. Third, we evaluate 13 state-of-the-art closed and open source LLMs on two educational datasets, OULAD (real assessment trajectories from the UK Open University) and PREDACT-CS (60 courses with real final grade outcomes and synthetically generated weekly score trajectories), alongside a human study with instructors and teaching assistants. We find that when tools are noisy, SOTA models are supposed to provide visibility to teachers so that they do not over-rely on wrong suggestions or hallucinations, but current models fail to do that. We offer PREDACTBENCH to help build better LLMs as AI decision support systems to help teachers.
Abstract (translated)
URL
https://arxiv.org/abs/2608.02372
PDF
-
Fast and Accurate Quotation Attribution in Literary Texts
2026-08-03 15:08:06Gaspard Michel, Hugo Attali, Elena V. EpureAbstract
Attributing quotations to their speakers in literary texts remains an open challenge. Standard methods, which independently predict a speaker mention for each quotation, are efficient but still limited in accuracy. In contrast, large language model (LLM) approaches achieve strong performance, but their computational cost limits their use in large-scale literary analysis. We propose an encoder-based efficient formulation that resolves multiple quotation attributions within a shared, large context window. Using our new formulation, \textit{joint scoring}, we report state-of-the-art (SOTA) performance on the Project Dialogism Novel Corpus (PDNC), comprising more than 35,000 manually annotated quotations from 22 English novels. Our best model reaches 94.5\% overall attribution accuracy while processing novels $20\times$ faster than comparable standard methods and more than $1000\times$ faster than LLM-based approaches on an A100 GPU. An analysis of models' representations suggests that joint scoring improves on challenging attribution examples by preserving long-range anaphora resolution signal, an information that we found already present in pretrained encoders. To facilitate adoption, we release ModernBookNLP, a modified fork of BookNLP that replaces its quotation attribution model with our best system available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2608.02359
PDF
-
CompanionBench: A Theory-Anchored, Real-World-Grounded Benchmark for AI Emotional Companionship
2026-08-03 10:44:04Yao Liu, Guangjia Chai, Yuming Huang, Jihao Huang, Lei Wang, Junchen WanAbstract
LLM companions are deployed at scale in personally consequential settings, yet poorly evaluated. Existing benchmarks use hand-authored scenarios and prompted simulators, aggregate empathy into one score, and overlook judge biases such as same-family favoritism and scale drift. We introduce CompanionBench, an interactive bilingual benchmark. To our knowledge, it is the first companion benchmark to ground both its scenarios and a trained user simulator in de-identified real-world data. A hidden disclosure gate branches each persona's trajectory on the agent's own behavior, controlling the interaction state space without scripting dialogue. We operationalize ten capabilities derived from 25 theories across psychology and counseling, four of them not graded explicitly by prior work: holding ambiguity, selfobject responsiveness, positive resonance and calibrated challenge. Agents are assessed on two complementary axes: a subjective ten-capability rubric and a deterministic measure of whether deeper disclosure was earned. A cross-family panel dilutes same-family favoritism; an Item Response Theory model separates agent quality from judge severity. Theory fixes what to measure and how personas are structured; real data supply events, history, and profiles -- coverage from theory, authenticity from data. Rankings are reproducible in both languages (rho = 0.996 ZH / 0.953 EN). Evaluating 28 agents reveals capability-level differences obscured by aggregate scores. Emotion regulation and calibrated challenge remain common weaknesses; holding ambiguity discriminates most. Role-play agents rank near the bottom: immersion does not imply relational competence. Across agents, the dominant failure mode is substituting surface warmth for substantive relational support. We will release 500 Chinese-English parallel pairs and the evaluation code.
Abstract (translated)
URL
https://arxiv.org/abs/2608.02046
PDF
-
PGMem: Tightly Coupled Persona-Memory Graph for Lifelong Personalized Agents
2026-08-03 05:14:59Wonjun Choi, Yerim Kim, Yukyung Lee, Susik YoonAbstract
Long-term personalized dialogue agents must track user preferences as their personas evolve. Existing memory systems organize past events well, but store personas as flat profiles detached from the events that justify them. This loose coupling leads to the memory-persona validity gap and the persona-aware retrieval gap. We propose PGMem, a heterogeneous persona-memory graph that connects event and persona nodes through typed provenance and evidence edges, keeping each persona signal traceable to the events that support or revise it. At retrieval time, PGMem expands from query-relevant seeds and ranks signals by evidential validity. Across three benchmarks with small language model backbones, PGMem consistently outperforms summary-based, persona-aware, graph-structured, and agentic memory baselines, and improves performance as the context grows. The source code of PGMem is available at this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2608.01708
PDF
-
Language Equality has a Price: A Systematic Investigation of Multi-turn LLM Performance for EU-24+
2026-08-02 17:20:01Sherzod Hakimov, Karl Osswald, Jelle Psurek, Eszter Bukovszky, A. Altar L\"user, David SchlangenAbstract
We evaluate large language models (LLMs) as language agents playing goal-directed dialogue games in self-play across 30 languages: the 24 official EU languages plus six others. Unlike static or preference-based evaluation, this paradigm is multi-turn, reference-free and programmatically scored, and because the game mechanics are language-agnostic it extends to a new language by localising a fixed set of prompt and word-list files. Evaluating nine open-weight and commercial LLMs, we find that no open-weight model covers the EU-24 well: in every official language both commercial systems outscore every open-weight model, and the two weakest average below 40 points across the EU-24. The commercial systems stay ahead even in languages with four orders of magnitude less public web text, showing that linguistic parity is achievable, but not from public crawls alone. A model's home region lifts it without closing the gap: Chinese is the strongest of all 30 languages for two Chinese-developed models, yet the best Chinese score of any model belongs to a US commercial system. Coverage is also not parity of service. Pooled over models and languages, the median non-English language costs 31% more to run than English, and scores 10% lower.
Abstract (translated)
URL
https://arxiv.org/abs/2608.01395
PDF
-
ShiJianBench: From Dialogue to Decision for Long-Horizon Evaluation of Investment Advisors
2026-08-02 12:41:47Jie Gong, Maowei Jiang, Zhiwei Liu, Yang Qiao, Wenxi Wu, Mengxi Xiao, Enze Zhang, Ziyan Kuang, Yankai Chen, Caishuang Huang, Meng Zhou, Xiku Du, Xue Liu, Guojun Xiong, Min Peng, Qianqian Xie, Sophia AnaniadouAbstract
Conversational investment advisors influence not only what users know, but also how they make subsequent decisions as market conditions evolve. Existing evaluations primarily assess response quality or observed outcomes, leaving the long-horizon pathway from advisor language to investor behavior difficult to audit. We introduce ShiJianBench, an offline framework for evaluating conversational investment advisors through matched investor trajectories under fixed historical market feedback. At its core is a multi-agent investor simulator with explicit evolving state variables, motive-driven deliberation, long-term memory, and dialogue-grounded updates. The simulator is calibrated against aggregate behavioral patterns from 7,199 real users, and advisor policies are evaluated using separate investor-side, service-side, and content-side metrics under a hard compliance gate. Experiments on Chinese fund-market traces from 2021 to 2026 identify a stable leading group of LLM advisors that combines substantially stronger personalized content with competitive investor-side trajectory outcomes. These results reveal a systematic distinction between producing a high-quality response and delivering an effective long-horizon intervention, motivating trajectory-aware evaluation of conversational advisors.
Abstract (translated)
URL
https://arxiv.org/abs/2608.01204
PDF
-
FraudBench: Stress-Testing Policy-Grounded Banking Agents Against Adaptive Fraud
2026-08-02 10:04:22Dheeraj Mohandas Pai, Lu XianAbstract
Conversational agents now act for end users through tools while holding access to customer databases and internal policy documents that a caller can reach through dialogue alone. Banking is the clearest case: the same agent that answers a question can also change contact details, reset a PIN, or move money, so ordinary customer service is inseparable from authorization, fraud detection, and policy compliance. Existing financial-fraud benchmarks classify static transactions or messages, and general agent-safety benchmarks target prompt injection or generic harmful use; none test whether a policy-grounded banking agent safely acts when a caller manipulates identity, authorization, and trust over a conversation. We introduce FraudBench, an executable benchmark built on the $\tau^2$-bench dual-control framework and the $\tau$-Knowledge banking environment. Both the agent and the simulated caller act through tools over shared, mutable account state, and the agent may grant the caller access to selected tools; the environment exposes a 698-document internal policy corpus that the agent must retrieve from. FraudBench contains 150 authored adversarial scenarios; a frozen public set of 107 (90 across ten fraud mechanisms plus 17 chained adaptive attacks) is used for all reported runs, with 43 further chained attacks held out. Safety is history-dependent: single-control tasks satisfy every precondition but one, and adaptive attacks make a later, locally valid request unsafe because of an earlier probe, admission, or failed attempt. Each scenario is annotated with observable evidence, prohibited actions, safe dispositions, and intervention points. A preliminary single-trial evaluation of four agents on the 107 graded tasks yields attack-security between 49\% and 65\%, with money-mule and first-party fraud the most common cross-model weaknesses.
Abstract (translated)
URL
https://arxiv.org/abs/2608.18136
PDF
-
JoyAI-Talker: Full-Duplex Speech Interactive Large Model Built for Empathetic Voice Agents
2026-08-02 09:27:51Yinhao Bai, Jinming Chen, Yafeng Chen, Wei Deng, Boya Dong, Nan Duan, Yu Gu, Weisheng Han, Yankun Huang, Ming Ke, Hao Li, Jingdong Li, Xiangyu Liang, Ning Liu, Yuan Liu, Ji Miao, Jiaqi Wang, Qi Wang, Wenchao Wang, Yuxuan Wang, Zhenfang Wang, Zhangyu Xiao, Chao Xue, Hongfei Xue, Fan Yu, Tianyi Zhang, Yuan Zhang, Yuqi Zhang, Lin ZhuAbstract
We present JoyAI-Talker, a full-duplex speech dialogue system that delivers robust foundation model capabilities while empowering empathetic interaction and voice agent intelligence. JoyAI-Talker adopts a modular Thinker-Talker architecture and further implements a unified speech-text joint training pipeline to mitigate the common "cognitive degradation" bottleneck, thereby largely preserving the model's core textual reasoning, STEM, and logical capabilities while extending them to speech-based interaction. For expressive speech synthesis, the Talker module employs a text-controllable generation paradigm that enables natural-language instructions to flexibly control vocal attributes and localized paralinguistic events, such as laughter and sighs, supporting more expressive and fine-grained speech responses. To enhance conversational empathy, we introduce the Persona-Adaptive Empathetic Response (PAER) framework. PAER employs a hierarchical cognitive pipeline to extract non-verbal speaker cues, such as gender, age, and emotional state, from raw input audio, incorporate them into the Thinker's CoT reasoning, and generate context-adaptive responses that align semantically appropriate text with fine-grained control over utterance-level expressiveness and localized paralinguistic events, including sighs, speaking rate, and volume. We further integrate Joy-Duplex, a state-driven, plug-and-play full-duplex framework that functions as an efficient gating engine for real-time turn control. Extensive evaluations show that JoyAI-Talker achieves highly competitive performance on foundational T2T and S2T benchmarks. In full-duplex evaluation, the system reaches a high response rate of 0.88 under user interruptions while maintaining an extremely low false-trigger rate under background speech, demonstrating its readiness for fluid and natural speech dialogue.
Abstract (translated)
URL
https://arxiv.org/abs/2608.01119
PDF
-
SoniSpeech: A Large-Scale Open-Vocabulary Tri-Modal Dataset for Wearable Silent Speech Interfaces
2026-08-01 18:03:37Ruidong Zhang, Jiacheng Liu, Fran\c{c}ois Guimbreti\`ere, Cheng ZhangAbstract
Wearable silent speech interfaces (SSIs) are limited to small, closed vocabularies. Approaches achieving larger vocabularies require obtrusive hardware such as facial electrodes. We present SoniSpeech, the first large-scale, open-vocabulary, trimodal dataset for wearable SSI using acoustic-sensing eyewear. It contains 34 hours across 18,000 utterances with three synchronized modalities: ultrasound echo profiles, voiced audio, and frontal video, in both voiced and silent modes. The corpus draws from the SODA dialogue dataset, providing contemporary conversational English with 5,356 unique words and full phoneme coverage. A CTC-based ResNet-34 baseline achieves 26.3% word error rate (WER) on open-vocabulary silent speech recognition, the first benchmark for this task. Dataset is available at this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2608.00803
PDF
-
DE-NER : Zero-shot Named Entity Recognition via Dialogue Elicitation of Large Language Models
2026-08-01 08:46:53Xuankang Zhang, Jiangming LiuAbstract
Recent advancements of zero-shot Named Entity Recognition (NER) establish strong baselines by formulating sequence labeling into question answering where Large Language Models (LLMs) can be naturally adopted. However, existing LLM-based zero-shot NER methods suffer from the limitations of prompt and demonstration engineering. To address these issues with minimal human interventions, we introduce DE-NER, a dialogue elicitation framework which elicits the chatting ability of LLMs to fully extract the knowledge encoded in LLMs. Our experiments demonstrate that the proposed method outperform the competitive baselines in zero-shot settings across multiple benchmarks, with an average improvement of 3.75\% F1 points. Codes are released in this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2608.00538
PDF
-
Bridging the English-Arabic Medical Knowledge Gap: Targeted Low-Rank Adaptation via Causal Layer Selection
2026-07-31 18:43:07Chaimae Abouzahir, Musa Khan, Hala Ali-Hassan, Congbo Ma, Khaled Saleh, Yousra Sadqi, Jihad Mallat, Walid Al-Eisawi, Nizar Habash, Farah E. ShamoutAbstract
Large Language Models (LLMs) perform strongly in English medical tasks but degrade substantially in Arabic, a gap widely attributed to limited training data. We systematically investigate this assumption via tuned lens probing and causal activation patching, and find that Arabic medical knowledge is present in intermediate model representations but fails to surface at the output. This mechanistic insight motivates a targeted adaptation strategy: rather than fine-tuning the full network, we propose Targeted Low-Rank Adaptation (TLoRA), restricted to the layer window where cross-lingual representations diverge, upstream of the output layers where the failure manifests. We evaluate TLoRA on multiple-choice medical QA, where our approach outperforms full-network LoRA, zero-shot, and few-shot baselines. We further evaluate it on short-answer generation and multi-turn clinical dialogue, where it performs competitively without the need for task-specific finetuning. We additionally introduce AraClinicDialog, a clinician-constructed Arabic medical dialogue benchmark in MSA with validated variants across four Arabic dialects. Together, these contributions demonstrate that mechanistic diagnosis can serve as a practical guide for targeted adaptation in underrepresented-language medical LLMs.
Abstract (translated)
URL
https://arxiv.org/abs/2608.00207
PDF
-
Sycophancy Undermines Epistemic Vigilance in Cooperative Vision-Language Tasks
2026-07-31 16:09:23Rupak Sarkar, Neha Srikanth, Saloni Gupta, Claire Bonial, Philip Resnik, Rachel RudingerAbstract
To maintain common ground in cooperative conversation, humans iteratively update their beliefs as conversation participants share new information; participants who are epistemically vigilant detect when new information conflicts with prior beliefs and take steps to repair these conflicts. In order for AI systems to serve as reliable partners in complex cooperative tasks, they must similarly weigh incoming information against their own private evidence and shared context and appropriately surface inconsistencies when they arise. To measure the epistemic vigilance of vision-language models in cooperative settings, we present an information-asymmetric, dialog-based "spot-the-difference" task. Two models are privately shown one image each, and must determine through conversation whether the images are identical or, if not, identify the difference. Models routinely fail at this: they frequently overlook key evidence in their private image in favor of agreeing with their conversational partner, even when their agreement is unwarranted. We relate these violations of epistemic vigilance to the broader behavior of sycophancy, which manifests itself in cooperative goal-oriented dialog as over-accommodation and weak evidential grounding. Our results show that model steering to reduce sycophancy with a vector learned from task-agnostic sycophancy examples can reduce epistemic vigilance-related errors, making models more faithful reporters of their evidence, and in turn, more reliable partners in information-asymmetric cooperative tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2607.29585
PDF
-
Beyond Retrieval: Analytic Memory for Multimodal Agents
2026-07-31 14:07:12Zhoujin Tian, Yao Tian, Hao Zhang, Cheng Chen, Yakun Li, Lei Zhang, Xiaofang ZhouAbstract
Long-term multimodal memory must support not only retrieving relevant information but also computing over observations accumulated across interactions. Existing systems largely emphasize \emph{retrieval memory}, organizing interaction histories through summaries and indexes to return query-relevant information at multiple granularities, from high-level abstractions to underlying records. In this paper, we formulate \emph{analytic memory} as a complementary abstraction that organizes recurring multimodal observations into queryable structures supporting filtering, aggregation, ranking, and temporal comparison. We present AdaMM, a framework that jointly supports retrieval and analytic memory. Rather than relying on application-defined schemas, AdaMM extracts provenance-linked attribute-value observations from dialogue, images, and contextual metadata, discovers recurring field structures, and materializes them for analytical access. At inference time, a memory-aware planner decomposes queries into retrieval and analytic operations and routes each operation to the appropriate tools. Experiments on two long-term multimodal memory benchmarks, MemEye and MemGallery, show that AdaMM improves performance by up to 11.3\% and 7.3\%, respectively.
Abstract (translated)
URL
https://arxiv.org/abs/2607.29440
PDF
-
Hy-MultiTurn: A Six-Dimensional Benchmark for Deep Multi-Turn Dialogue Understanding
2026-07-31 09:15:56Eileen Ye, Jiawen Tao, Yaoming Li, Chenxu Liu, Wenhan Yu, Yaxin Fan, Xiaokun Yuan, Mengzhou Wu, Yanbing Jiang, Maxm PanAbstract
Long-running multi-turn interactions with chatbots and agents are now common, and a correct response often depends on remembering earlier details, tracking later revisions, identifying intended objects or referents, and withholding action when required conditions are unmet. Existing multi-turn benchmarks typically cover short exchanges and do not fully evaluate these capabilities in long multi-turn interactions, particularly in Chinese, while offering limited insight into how and why models fail. To address these limitations, we analyze real chatbot failures to identify six recurring mechanisms and use them to define six controlled evaluation modes in Hy-MultiTurn, a Chinese benchmark for deep multi-turn dialogue understanding. The six modes evaluate constraint memory, precise execution, constraint synthesis, object localization, action suppression, and reference resolution. Across the six modes, we construct 209 controlled tasks spanning 12-76 turns, with dialogue length, irrelevant-topic distraction, and colloquial phrasing adding further difficulty. Evaluation of 22 frontier model configurations shows that Hy-MultiTurn is broadly challenging, as even GPT-5.5, the strongest overall configuration, satisfies all requirements in only 41.1 percent of responses and no model performs best in all six modes.
Abstract (translated)
URL
https://arxiv.org/abs/2607.29196
PDF
-
M3-DuplexBench: A Multi-Turn, Multilingual, Multidomain Benchmark for Full-Duplex Spoken Dialogue Models
2026-07-31 07:56:25Ryo Fukuda, Atsushi Ando, Hiroki Kanagawa, Takatomo Kano, Marc Delcroix, Naohiro Tawara, Yuya ChibaAbstract
Full-duplex spoken dialogue systems (FDSDSs) can listen while speaking, enabling natural behaviors such as smooth turn-taking, backchannel handling, and user barge-in handling. However, fair comparisons in multi-turn conversations remain a challenge. In addition, existing benchmarks provide limited coverage of languages and dialogue domains. We propose M3-DuplexBench, a multi-turn, multilingual, multidomain benchmark for FDSDSs. M3-DuplexBench supports English and Japanese and covers both casual conversation and multi-turn question answering. In addition, we evaluate models under multiple dialogue context settings, including single-turn, user-only, and teacher-forced full-context settings, to analyze how dialogue history affects model behavior. Experiments with recent FDSDSs reveal model-specific turn-taking characteristics, clear performance gaps across languages and domains, and mixed effects of dialogue context.
Abstract (translated)
URL
https://arxiv.org/abs/2607.29125
PDF
-
TRACE Bench: Task-driven Roleplay Agentic Checklist Evaluation
2026-07-31 07:35:59Jiahui Zhang, Ziwei Zhang, Yipeng Wang, Yibo Liu, Haozhou Pang, Yikai Hu, Hongyan Ren, Lan Zhou, Qi Gan, Kai ShengAbstract
Roleplay evaluation should do more than assign a single score: it should reveal which role requirements were tested, which failed, and which dialogue evidence supports the judgment. We propose TRACE Bench, a task-driven agentic checklist evaluation framework. It decomposes each role profile offline into a fixed checklist, then uses a User Agent to converse naturally with the target roleplay model while privately updating checklist states from model responses. Scores therefore trace back to checklist items and supporting dialogue turns rather than a black-box holistic impression. For coverage cross-validation, we audit released M2 free-dialogue transcripts from the MiniMax Role-play Benchmark against the same role-derived checklist. The released free-chat transcripts cover only 73.74% of key role-profile points, whereas TRACE Bench reaches 99.91% coverage in fewer turns. Robustness experiments show stable rankings under repeated runs and User Agent replacement. Across 26 models, TRACE Bench reports overall rankings together with capability breakdowns and checklist traces. It also supports Closed-Loop Benchmark Evolution, distilling verification methods proven effective in failed traces so later evaluations can more reliably elicit and examine observed failure modes.
Abstract (translated)
URL
https://arxiv.org/abs/2608.11236
PDF
-
MMShopBench: A Real-Log Benchmark for Multimodal, Multi-Turn Shopping Agents
2026-07-31 04:00:15Zeying Hao, Hao Guo, Mengtao Xu, Yimin Hu, Yuheng Song, Zesheng Zhou, Jinsong Lan, Xiaoyong ZhuAbstract
Online shoppers increasingly turn to AI shopping assistants, using images and multi-turn dialogue to express and refine product needs that are difficult to articulate in text alone. However, existing benchmarks largely rely on text-only or synthetic requests, underrepresenting complex real-world shopping requirements jointly expressed through images and language. We introduce MMShopBench, the first real-log benchmark for multimodal, multi-turn shopping agents. Built from carefully cleaned and manually annotated shopping logs, MMShopBench provides ground-truth annotations of each request's purchase intent and mandatory product requirements. Agents must infer these requirements jointly from user images and multi-turn dialogue, retrieve candidate products through image and text search, and verify that each candidate satisfies all requirements using its product images and structured attributes. We evaluate representative open-source and proprietary models using an evidence-grounded multimodal protocol and construct a companion training set for fine-tuning an open-source model. To ensure reproducible experimentation, we build an offline shopping sandbox, where fine-tuning substantially narrows the performance gap between our open-source model and leading proprietary models, demonstrating the effectiveness of our training data.
Abstract (translated)
URL
https://arxiv.org/abs/2607.29002
PDF
