This work introduces WebXSkill, a framework that bridges a grounding gap with executable skills, each pairing a parameterized action program with step-level natural-language guidance, and finds that better skill deployment mode depends on a model's plan and execution capability.
A trajectory-aware evolutionary search method, T-MAP, which leverages execution trajectories to guide the discovery of adversarial prompts and enables the automatic generation of attacks that not only bypass safety guardrails but also reliably realize harmful objectives through actual tool interactions.
Hyomin Lee, Sangwoo Park, Yumin Choi et al.· arXiv.org· 8 citations· ⚡1
Simple episodic retrieval is established as a strong foundation for agent memory and retrieval-aware fine-tuning as a practical and effective framework for building agents that learn to learn from experience.
Thomas Palmeira Ferraz, Romain Deffayet, Vassilina Nikoulina et al.· arXiv.org· 6 citations
SHARD-MEMO is presented, an agentic memory system built on scope-before-routing: metadata predicates first identify the admissible shards, and a learned router then selects a small number of them for shard-local approximate nearest neighbor retrieval, separating hard admissibility from learned relevance ranking.
Yang Zhao, Chengxiao Dai, Mengyi Kou et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
It is found that while most models exhibit significant sycophantic tendencies in the common setting, seven of the models exhibit ``moral remorse'', five of which significantly over-compensate for their sycophancy in case it explicitly harms a third party.
Shahar Ben Natan, Oren Tsur· arXiv.org· 1 citation
SPRInG is introduced, a novel semi-parametric framework designed for effective continual personalization that significantly outperforms existing baselines, validating its robustness for real-world continual personalization.
This work introduces DEER-3D, an error-driven framework that diagnoses grounding failures and generates targeted counterfactual training supervision via a structured "Decompose, Diagnose, Edit, and Retrain"loop, and underscores the effectiveness of targeted, error-driven scene editing in bridging linguistic reasoning with spatial grounding in 3D LLMs.
Yue Zhang, Zun Wang, Han Lin et al.· arXiv.org· 0 citations
Two test-time monitoring strategies are proposed: reasoning expression monitoring and hidden states monitoring, that reduce token usage by 62.7-93.6%, substantially improving efficiency and reliability while largely preserving accuracy.
Qingjie Zhang, Yujia Fu, Yang Wang et al.· 2 citations· ⚡1
Persuasion Duality is identified: reasoning enhances an agent's persuasive power while simultaneously increasing its resistance to persuasion, and a prompt-level adversarial argument detection method is proposed that consistently improves agent robustness.
D2Snap, an algorithm to downsample the DOM, premised on preserving actionability and actionability-discriminating features, is proposed and evaluated using a snapshot-variant web agent (GPT-4o) on a dataset sampled from Online-Mind2Web.
Thassilo M. Schiepanski, Nicholas Piël· arXiv.org· 4 citations
In psychological counseling, effective support is not always delivered through long, information-rich responses. Minimal responses, such as backchannel cues and concise empathic statements, help convey attentive listening, express empathy, and encourage clients to continue expressing themselves. However, existing counseling dialogue systems and evaluation frameworks often favor explicit, content-rich replies, overlooking the interactional value of brief counselor utterances. This paper presents a systematic cross-lingual analysis of minimal responses across multiple counseling dialogue datasets. We develop a two-stage filtering method based on utterance length and content, followed by contextual verification using a large language model (LLM). Our analysis shows that minimal responses are common in human-collected datasets but substantially underrepresented in LLM-generated ones. We further evaluate current LLMs in manually curated dialogue contexts where human counselors used minimal responses. The results show that strong commercial LLMs are capable of generating minimal responses when explicitly instructed, but still struggle to determine when such responses are appropriate. Counseling-specific models trained on synthetic data perform particularly poorly, tending instead to produce longer and more information-rich responses. Moreover, LLM-based response-quality evaluation may undervalue minimal responses, even when they are interactionally appropriate.
Inspired by the Big Five model's dimensional view of personality, a framework that reframes authorial writing characteristics as coordinates within a unified and interpretable space is proposed, which improves authorial expressiveness while preserving semantic fidelity.
Jinghui Zhang, Lang Gao, Ao Li et al.· 0 citations
