Multimodal long-chain-of-thought (CoT) generation requires models to dynamically route information between visual evidence and an expanding reasoning history. Existing hallucination mitigation methods often strengthen visual grounding or intervene in decoding, but typically treat grounding as a relatively static property of generation. We propose a training-free dynamic attention intervention framework that exploits the head-level sparsity, layer-level stratification, and trajectory-level dynamics of multimodal long-CoT decoding. Our method selects routing-sensitive heads using perturbation-only calibration, assigns visual-evidence and reasoning-history interventions to different layer regions, and continuously rescales selected head outputs according to their instantaneous token-group attention ratios. Across multiple MLLM architectures and multimodal hallucination benchmarks, our method consistently improves over vanilla decoding and visual contrastive decoding, achieving an average performance gain of 4.6%.

Reinforcement learning (RL) has shown promise for training agentic models that engage in dynamic, multi-turn interactions. However, the value of such agents ultimately depends on their ability to assist users with diverse goals, preferences, and evolving interaction patterns. We introduce **UserRL**, a reproducible and extensible testbed for training and evaluating user-centric agents through standardized gym environments paired with simulated users. UserRL supports flexible reward specification by decoupling turn-level reward shaping from trajectory-level scoring, enabling controlled studies of credit assignment in multi-turn RL. Across Qwen3 models, our experiments show that SFT cold start provides an effective initialization for user-centric RL; trajectory-level reward-to-go scoring improves both effectiveness and interaction efficiency; and stronger simulated users improve training, while open-source simulators remain a cost-effective alternative with positive cross-simulator and preliminary real-user transfer. These results highlight reward design and user simulation as key to user-centric agent training, positioning UserRL as a practical foundation for interactive agent RL.

Large Language Model agents increasingly solve tool-use tasks, yet often fail to determine what users actually want when goals are underspecified, incremental, or indirectly expressed. We introduce **UserBench**, a controlled Gymnasium environment for evaluating proactive preference elicitation in multi-turn, tool-grounded interactions. Instantiated in travel planning, UserBench hides user preferences behind implicit natural-language cues and requires agents to ask targeted clarification questions, search structured option databases, and select user-aligned options. Evaluating leading open- and closed-source models reveals a large gap between tool execution and user understanding: models achieve high valid-search rates but elicit only a minority of hidden preferences, and even stronger models remain far below a human oracle. Further analysis shows that many additional interaction turns are broad or non-informative rather than preference-relevant. UserBench provides a reproducible testbed for studying and training user-centric agent interaction while intentionally abstracting away full travel negotiation and infeasible-constraint recovery.

Autonomous agents powered by large language models (LLMs) have attracted significant research interest. However, the open-source community faces many challenges in developing specialized models for agent tasks, driven by the scarcity of high-quality agent datasets and the absence of standard protocols in this area. We introduce xLAM, a series of large action models designed for AI agent tasks. The xLAM series includes five models with both dense and mixture-of-expert architectures, ranging from 1B to 8x22B parameters, trained using a scalable, flexible pipeline that unifies, augments, and synthesizes diverse datasets to enhance AI agents’ generalizability and performance across varied environments. Our experimental results demonstrate that xLAM consistently delivers exceptional performance across multiple agent ability benchmarks, notably securing the 1st position on the Berkeley Function-Calling Leaderboard, outperforming GPT-4, Claude-3, and many other models in terms of tool use. By releasing the xLAM series, we aim to advance the performance of open-source LLMs for autonomous AI agents, potentially accelerating progress and democratizing access to high-performance models for agent tasks.

Personalization is essential for AI assistants, especially in private AI settings where models are expected to interpret users’ personal data (e.g., conversations, app usage) to understand their background, preferences, and social context. However, due to privacy concerns, existing academic research lacks direct access to such data, making benchmarking difficult. To fill this gap, we propose a synthetic data pipeline that generates realistic user profiles and private documents, enabling the creation of PersonaBench—a benchmark for evaluating models’ ability to understand personal information. Using this benchmark, we assess Retrieval-Augmented Generation (RAG) pipelines on personalized questions and find that current models struggle to accurately extract and answer questions even when provided with the full set of user documents, highlighting the need for improved personalization methods.

Large Action models are essential for enabling autonomous agents to perform complex tasks. However, training such models remains challenging due to the diversity of agent environments and the complexity of noisy agentic data. Existing infrastructure offers limited support for scalable, agent-specific fine-tuning and standardized agent data processing. We introduce ActionStudio, a lightweight and extensible data and training framework designed for large action models. ActionStudio unifies diverse agent trajectories using our proposed Unified Format 2.0, supports a range of training workflows with optimized multi-node distributed setup, and integrates robust preprocessing and real-time verification tools. ActionStudio demonstrates up to 9× higher throughput compared to existing agentic training frameworks, and our trained models yield top performances across public and realistic agent benchmarks. To support the broader research community, we open-source the ActionStudio framework and release actionstudio-98k, a curated dataset of 98k high-quality trajectories.

In today’s rapidly evolving business landscape, organizations are turning to AI agents to automate tasks, streamline business operations, and improve decision-making processes. However, despite the flexibility offered by existing libraries, the developed agents often struggle with integration into organizational workflows, resulting in limited daily usage for work. In this paper, we present SlackAgents, a multi-agent library for scalable management and collaboration of AI agents on Slack. As an agentic layer developed upon the Slack platform, the framework offers instant AI integration into organizational workflows and enables AI-powered automation of real daily tasks. Furthermore, SLACKAGENTS facilitates scalable collaboration, allowing for effective communication and task orchestration. Our solution bridges existing gaps, offering a robust platform for developing, deploying and managing AI agents for workplace environments.

The rapid adoption of Large Language Models (LLMs) as intelligent agents has underscored the necessity for robust evaluation frameworks capable of assessing agent performance in realistic, interactive environments. Existing evaluation methodologies often suffer from limitations such as static task benchmarks, limited scope, and inadequate integration with practical applications. In response, we introduce MCPEval, an open-source, Model Context Protocol (MCP)-based evaluation framework specifically tailored for comprehensive and systematic assessment of LLM-powered agents. MCPEval standardizes evaluations across diverse domains through automated task generation and verification, supports multiple performance metrics, and integrates seamlessly with native agent capabilities. We empirically validate the effectiveness of MCPEval across five distinct real-world domains, highlighting significant variations in performance across various LLM architectures and prompting strategies. Our results illustrate the framework’s capacity to uncover nuanced performance patterns and identify domain-specific strengths and weaknesses, providing valuable insights beyond traditional binary success metrics. We publicly release MCPEval to foster reproducible research and promote standardized evaluation practices within the LLM agent community.

We introduce the Principled Reasoning and Acting (PRAct) framework, a novel method for learning and enforcing action principles from trajectory data. Central to our approach is the use of text gradients from a reflection and optimization engine to derive these action principles. To adapt action principles to specific task requirements, we propose a new optimization framework, Reflective Principle Optimization (RPO). After execution, RPO employs a reflector to critique current action principles and an optimizer to update them accordingly.We investigate the RPO framework under two scenarios: Reward-RPO, which uses environmental rewards for reflection, and Self-RPO, which conducts self-reflection without external rewards. Additionally, we developed two RPO methods, RPO-Traj and RPO-Batch, to adapt to different settings.Experimental results across four environments demonstrate that the PRAct agent, leveraging the RPO framework, can effectively learn and apply action principles to enhance performance.