AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning

Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving frontier. Like human cognitive development, agents are expected to acquire knowledge and skills through exploration and interaction with the environment. Despite advances, the community still lacks a unified, interactive reinforcement learning (RL) framework that can effectively train such agents from scratch -- without relying on supervised fine-tuning (SFT) -- across diverse and realistic environments. To bridge this gap, we introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL. The framework features a modular and decoupled architecture, ensuring high flexibility and extensibility. It encompasses a wide variety of real-world scenarios, and supports mainstream RL algorithms. Furthermore, we propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization. In early stages, it emphasizes exploitation by restricting the number of interactions, and gradually shifts towards exploration with larger horizons to encourage diverse problem-solving strategies. In this way, the agent develops more diverse behaviors and is less prone to collapse under long horizons. We perform extensive experiments to validate the stability and effectiveness of both the AgentGym-RL framework and the ScalingInter-RL approach. Our agents match or surpass commercial models on 27 tasks across diverse environments. We offer key insights and will open-source the complete AgentGym-RL framework -- including code and datasets -- to empower the research community in developing the next generation of intelligent agents.

A Knowledge-Driven Diffusion Policy for End-to-End Autonomous Driving Based on Expert Routing

End-to-end autonomous driving remains constrained by the need to generate multi-modal actions, maintain temporal stability, and generalize across diverse scenarios. Existing methods often collapse multi-modality, struggle with long-horizon consistency, or lack modular adaptability. This paper presents KDP, a knowledge-driven diffusion policy that integrates generative diffusion modeling with a sparse mixture-of-experts routing mechanism. The diffusion component generates temporally coherent and multi-modal action sequences, while the expert routing mechanism activates specialized and reusable experts according to context, enabling modular knowledge composition. Extensive experiments across representative driving scenarios demonstrate that KDP achieves consistently higher success rates, reduced collision risk, and smoother control compared to prevailing paradigms. Ablation studies highlight the effectiveness of sparse expert activation and the Transformer backbone, and activation analyses reveal structured specialization and cross-scenario reuse of experts. These results establish diffusion with expert routing as a scalable and interpretable paradigm for knowledge-driven end-to-end autonomous driving.

A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers

Scientific Large Language Models (Sci-LLMs) are transforming how knowledge is represented, integrated, and applied in scientific research, yet their progress is shaped by the complex nature of scientific data. This survey presents a comprehensive, data-centric synthesis that reframes the development of Sci-LLMs as a co-evolution between models and their underlying data substrate. We formulate a unified taxonomy of scientific data and a hierarchical model of scientific knowledge, emphasizing the multimodal, cross-scale, and domain-specific challenges that differentiate scientific corpora from general natural language processing datasets. We systematically review recent Sci-LLMs, from general-purpose foundations to specialized models across diverse scientific disciplines, alongside an extensive analysis of over 270 pre-/post-training datasets, showing why Sci-LLMs pose distinct demands -- heterogeneous, multi-scale, uncertainty-laden corpora that require representations preserving domain invariance and enabling cross-modal reasoning. On evaluation, we examine over 190 benchmark datasets and trace a shift from static exams toward process- and discovery-oriented assessments with advanced evaluation protocols. These data-centric analyses highlight persistent issues in scientific data development and discuss emerging solutions involving semi-automated annotation pipelines and expert validation. Finally, we outline a paradigm shift toward closed-loop systems where autonomous agents based on Sci-LLMs actively experiment, validate, and contribute to a living, evolving knowledge base. Collectively, this work provides a roadmap for building trustworthy, continually evolving artificial intelligence (AI) systems that function as a true partner in accelerating scientific discovery.

Module-Aware Parameter-Efficient Machine Unlearning on Transformers

Transformer has become fundamental to a vast series of pre-trained large models that have achieved remarkable success across diverse applications. Machine unlearning, which focuses on efficiently removing specific data influences to comply with privacy regulations, shows promise in restricting updates to influence-critical parameters. However, existing parameter-efficient unlearning methods are largely devised in a module-oblivious manner, which tends to inaccurately identify these parameters and leads to inferior unlearning performance for Transformers. In this paper, we propose {\tt MAPE-Unlearn}, a module-aware parameter-efficient machine unlearning approach that uses a learnable pair of masks to pinpoint influence-critical parameters in the heads and filters of Transformers. The learning objective of these masks is derived by desiderata of unlearning and optimized through an efficient algorithm featured by a greedy search with a warm start. Extensive experiments on various Transformer models and datasets demonstrate the effectiveness and robustness of {\tt MAPE-Unlearn} for unlearning.

Mimicking the Physicist's Eye:A VLM-centric Approach for Physics Formula Discovery

Automated discovery of physical laws from observational data in the real world is a grand challenge in AI. Current methods, relying on symbolic regression or LLMs, are limited to uni-modal data and overlook the rich, visual phenomenological representations of motion that are indispensable to physicists. This "sensory deprivation" severely weakens their ability to interpret the inherent spatio-temporal patterns within dynamic phenomena. To address this gap, we propose VIPER-R1, a multimodal model that performs Visual Induction for Physics-based Equation Reasoning to discover fundamental symbolic formulas. It integrates visual perception, trajectory data, and symbolic reasoning to emulate the scientific discovery process. The model is trained via a curriculum of Motion Structure Induction (MSI), using supervised fine-tuning to interpret kinematic phase portraits and to construct hypotheses guided by a Causal Chain of Thought (C-CoT), followed by Reward-Guided Symbolic Calibration (RGSC) to refine the formula structure with reinforcement learning. During inference, the trained VIPER-R1 acts as an agent: it first posits a high-confidence symbolic ansatz, then proactively invokes an external symbolic regression tool to perform Symbolic Residual Realignment (SR^2). This final step, analogous to a physicist's perturbation analysis, reconciles the theoretical model with empirical data. To support this research, we introduce PhysSymbol, a new 5,000-instance multimodal corpus. Experiments show that VIPER-R1 consistently outperforms state-of-the-art VLM baselines in accuracy and interpretability, enabling more precise discovery of physical laws. Project page: https://jiaaqiliu.github.io/VIPER-R1/

Trade-offs in Image Generation: How Do Different Dimensions Interact?

Model performance in text-to-image (T2I) and image-to-image (I2I) generation often depends on multiple aspects, including quality, alignment, diversity, and robustness. However, models' complex trade-offs among these dimensions have rarely been explored due to (1) the lack of datasets that allow fine-grained quantification of these trade-offs, and (2) the use of a single metric for multiple dimensions. To bridge this gap, we introduce TRIG-Bench (Trade-offs in Image Generation), which spans 10 dimensions (Realism, Originality, Aesthetics, Content, Relation, Style, Knowledge, Ambiguity, Toxicity, and Bias), contains 40,200 samples, and covers 132 pairwise dimensional subsets. Furthermore, we develop TRIGScore, a VLM-as-judge metric that automatically adapts to various dimensions. Based on TRIG-Bench and TRIGScore, we evaluate 14 models across T2I and I2I tasks. In addition, we propose the Relation Recognition System to generate the Dimension Trade-off Map (DTM) that visualizes the trade-offs among model-specific capabilities. Our experiments demonstrate that DTM consistently provides a comprehensive understanding of the trade-offs between dimensions for each type of generative model. Notably, we show that the model's dimension-specific weaknesses can be mitigated through fine-tuning on DTM to enhance overall performance. Code is available at: https://github.com/fesvhtr/TRIG

Interactive Adversarial Testing of Autonomous Vehicles with Adjustable Confrontation Intensity

Scientific testing techniques are essential for ensuring the safe operation of autonomous vehicles (AVs), with high-risk, highly interactive scenarios being a primary focus. To address the limitations of existing testing methods, such as their heavy reliance on high-quality test data, weak interaction capabilities, and low adversarial robustness, this paper proposes ExamPPO, an interactive adversarial testing framework that enables scenario-adaptive and intensity-controllable evaluation of autonomous vehicles. The framework models the Surrounding Vehicle (SV) as an intelligent examiner, equipped with a multi-head attention-enhanced policy network, enabling context-sensitive and sustained behavioral interventions. A scalar confrontation factor is introduced to modulate the intensity of adversarial behaviors, allowing continuous, fine-grained adjustment of test difficulty. Coupled with structured evaluation metrics, ExamPPO systematically probes AV's robustness across diverse scenarios and strategies. Extensive experiments across multiple scenarios and AV strategies demonstrate that ExamPPO can effectively modulate adversarial behavior, expose decision-making weaknesses in tested AVs, and generalize across heterogeneous environments, thereby offering a unified and reproducible solution for evaluating the safety and intelligence of autonomous decision-making systems.

InfoDeepSeek: Benchmarking Agentic Information Seeking for Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by grounding responses with retrieved information. As an emerging paradigm, Agentic RAG further enhances this process by introducing autonomous LLM agents into the information seeking process. However, existing benchmarks fall short in evaluating such systems, as they are confined to a static retrieval environment with a fixed, limited corpus} and simple queries that fail to elicit agentic behavior. Moreover, their evaluation protocols assess information seeking effectiveness by pre-defined gold sets of documents, making them unsuitable for the open-ended and dynamic nature of real-world web environments. To bridge this gap, we present InfoDeepSeek, a new benchmark with challenging questions designed for assessing agentic information seeking in real-world, dynamic web environments. We propose a systematic methodology for constructing challenging queries satisfying the criteria of determinacy, difficulty, and diversity. Based on this, we develop the first evaluation framework tailored to dynamic agentic information seeking, including fine-grained metrics about the accuracy, utility, and compactness of information seeking outcomes. Through extensive experiments across LLMs, search engines, and question types, InfoDeepSeek reveals nuanced agent behaviors and offers actionable insights for future research.

Towards Human-Centric Autonomous Driving: A Fast-Slow Architecture Integrating Large Language Model Guidance with Reinforcement Learning

Autonomous driving has made significant strides through data-driven techniques, achieving robust performance in standardized tasks. However, existing methods frequently overlook user-specific preferences, offering limited scope for interaction and adaptation with users. To address these challenges, we propose a "fast-slow" decision-making framework that integrates a Large Language Model (LLM) for high-level instruction parsing with a Reinforcement Learning (RL) agent for low-level real-time decision. In this dual system, the LLM operates as the "slow" module, translating user directives into structured guidance, while the RL agent functions as the "fast" module, making time-critical maneuvers under stringent latency constraints. By decoupling high-level decision making from rapid control, our framework enables personalized user-centric operation while maintaining robust safety margins. Experimental evaluations across various driving scenarios demonstrate the effectiveness of our method. Compared to baseline algorithms, the proposed architecture not only reduces collision rates but also aligns driving behaviors more closely with user preferences, thereby achieving a human-centric mode. By integrating user guidance at the decision level and refining it with real-time control, our framework bridges the gap between individual passenger needs and the rigor required for safe, reliable driving in complex traffic environments.

MemeBLIP2: A novel lightweight multimodal system to detect harmful memes

Memes often merge visuals with brief text to share humor or opinions, yet some memes contain harmful messages such as hate speech. In this paper, we introduces MemeBLIP2, a light weight multimodal system that detects harmful memes by combining image and text features effectively. We build on previous studies by adding modules that align image and text representations into a shared space and fuse them for better classification. Using BLIP-2 as the core vision-language model, our system is evaluated on the PrideMM datasets. The results show that MemeBLIP2 can capture subtle cues in both modalities, even in cases with ironic or culturally specific content, thereby improving the detection of harmful material.