Legal general intelligence (GI) refers to artificial intelligence (AI) that encompasses legal understanding, reasoning, and decision-making, simulating the expertise of legal experts across domains. However, existing benchmarks are result-oriented and fail to systematically evaluate the legal intelligence of large language models (LLMs), hindering the development of legal GI. To address this, we propose LexGenius, an expert-level Chinese legal benchmark for evaluating legal GI in LLMs. It follows a Dimension-Task-Ability framework, covering seven dimensions, eleven tasks, and twenty abilities. We use the recent legal cases and exam questions to create multiple-choice questions with a combination of manual and LLM reviews to reduce data leakage risks, ensuring accuracy and reliability through multiple rounds of checks. We evaluate 12 state-of-the-art LLMs using LexGenius and conduct an in-depth analysis. We find significant disparities across legal intelligence abilities for LLMs, with even the best LLMs lagging behind human legal professionals. We believe LexGenius can assess the legal intelligence abilities of LLMs and enhance legal GI development.Our project is available at https://github.com/QwenQKing/LexGenius.

Large language model (LLM)-based multi-agent systems (MAS) have shown strong capabilities in solving complex tasks. As MAS become increasingly autonomous in various safety-critical tasks, detecting malicious agents has become a critical security concern. Although existing graph anomaly detection (GAD)-based defenses can identify anomalous agents, they mainly rely on coarse sentence-level information and overlook fine-grained lexical cues, leading to suboptimal performance. Moreover, the lack of interpretability in these methods limits their reliability and real-world applicability. To address these limitations, we propose , an explainable and fine-grained safeguarding framework for detecting malicious agents in MAS. To incorporate both coarse and fine-grained textual information for anomalous agent identification, we utilize a bi-level agent encoder to jointly model the sentence- and token-level representations of each agent. A theme-based anomaly detector further captures the evolving discussion focus in MAS dialogues, while a bi-level score fusion mechanism quantifies token-level contributions for explanation. Extensive experiments across diverse MAS topologies and attack scenarios demonstrate robust detection performance and strong interpretability of XG-Guard.

The security of LLM-based multi-agent systems (MAS) is critically threatened by propagation vulnerability, where malicious agents can distort collective decision-making through inter-agent interactions. While existing supervised defense methods demonstrate promising performance, they may be impractical in real-world scenarios due to their heavy reliance on labeled malicious agents to train a supervised malicious detection model. To enable practical and generalizable MAS defenses, in this paper, we propose BlindGuard, an unsupervised defense method that learns without requiring any attack-specific labels or prior knowledge of malicious behaviors. To this end, we establish a hierarchical agent encoder to capture individual, neighborhood, and global interaction patterns of each agent, providing a comprehensive understanding for malicious agent detection. Meanwhile, we design a corruption-guided detector that consists of directional noise injection and contrastive learning, allowing effective detection model training solely on normal agent behaviors. Extensive experiments show that BlindGuard effectively detects diverse attack types across MAS with various communication patterns while maintaining superior generalizability compared to supervised baselines.

LLM-based Multi-Agent Systems (MAS) have demonstrated remarkable capabilities in solving complex tasks. Central to MAS is the communication topology which governs how agents exchange information internally. Consequently, the security of communication topologies has attracted increasing attention. In this paper, we investigate a critical privacy risk: MAS communication topologies can be inferred under a restrictive black-box setting, exposing system vulnerabilities and posing significant intellectual property threats. To explore this risk, we propose Communication Inference Attack (CIA), a novel attack that constructs new adversarial queries to induce intermediate agents’ reasoning outputs and models their semantic correlations through the proposed global bias disentanglement and LLM-guided weak supervision. Extensive experiments on MAS with optimized communication topologies demonstrate the effectiveness of CIA, achieving an average AUC of 0.87 and a peak AUC of up to 0.99, thereby revealing the substantial privacy risk in MAS. The source code is available at https://github.com/aabbbcd/CIA.

Knowledge editing (KE) has recently emerged as a promising technique to update specific facts in large language models (LLMs) without full retraining. While existing KE methods show promising results on general-domain benchmarks, their effectiveness in the medical domain remains largely unexplored. Medical knowledge editing poses unique challenges, requiring models not only to memorize new facts but also to internalize and generalize them for reliable and interpretable clinical decision-making. In this work, we propose MedEditBench, a rigorous evaluation framework for assessing medical knowledge editing. Our preliminary results reveal that current KE paradigm, which directly edits simple answers to the LLMs, often leads to superficial updates with poor generalization. To address this, we introduce Self-Generated Rationale Editing (SGR-Edit), which leverages model-generated rationales as editing targets, enabling deeper knowledge integration. Extensive experiments across diverse LLMs and KE methods demonstrate that SGR-Edit consistently improves editing efficacy and generalization. Furthermore, we examine the impact of sequential edits on in-domain medical knowledge, external-domain knowledge, as well as general model capabilities, offering practical insights for deploying KE in real-world medical applications.

Agentic Reinforcement Learning (RL) shows promise for complex tasks, but Text-to-SQL remains mostly restricted to single-turn paradigms. A primary bottleneck is the credit assignment problem. In traditional paradigms, rewards are determined solely by the final-turn feedback, which ignores the intermediate process and leads to ambiguous credit evaluation. To address this, we propose Agentic SQL, a framework featuring a universal two-tiered reward mechanism designed to provide effective trajectory-level evaluation and dense step-level signals. First, we introduce Aggregated Trajectory Reward (ATR) to resolve multi-turn credit assignment. Using an asymmetric transition matrix, ATR aggregates process-oriented scores to incentivize continuous improvement. Leveraging Lyapunov stability theory, we prove ATR acts as an energy dissipation operator, guaranteeing a cycle-free policy and monotonic convergence. Second, Column-Set Matching Reward (CSMR) provides immediate step-level rewards to mitigate sparsity. By executing queries at each turn, CSMR converts binary (0/1) feedback into dense [0,1] signals based on partial correctness. Evaluations on BIRD show a 5% gain over binary-reward GRPO. Notably, our approach outperforms SOTA Arctic-Text2SQL-R1-7B on BIRD and Spider 2.0 using identical models, propelling Text-to-SQL toward a robust multi-turn agent paradigm.

The communication topology in large language model-based multi-agent systems fundamentally governs inter-agent collaboration patterns, critically shaping both the efficiency and effectiveness of collective decision-making. While recent studies for communication topology automated design tend to construct sparse structures for efficiency, they often overlook why and when sparse and dense topologies help or hinder collaboration. In this paper, we present a causal framework to analyze how agent outputs, whether correct or erroneous, propagate under topologies with varying sparsity. Our empirical studies reveal that moderately sparse topologies, which effectively suppress error propagation while preserving beneficial information diffusion, typically achieve optimal task performance. Guided by this insight, we propose a novel topology design approach, EIB-Learner, that balances error suppression and beneficial information propagation by fusing connectivity patterns from both dense and sparse graphs. Extensive experiments show the superior effectiveness, communication cost, and robustness of EIB-Learner.

Knowledge graph completion (KGC) aims to infer new knowledge and make predictions from knowledge graphs. Recently, large language models (LLMs) have exhibited remarkable reasoning capabilities. LLM-enhanced KGC methods primarily focus on designing task-specific instructions, achieving promising advancements. However, there are still two critical challenges. First, existing methods often ignore the inconsistent representation spaces between natural language and graph structures. Second, most approaches develop separate instructions for different KGC tasks, leading to duplicate works and time-consuming processes. To address these challenges, we propose SAT, a novel framework that enhances LLMs for KGC via structure-aware alignment-tuning. Specifically, we first introduce hierarchical knowledge alignment to align graph embeddings with the natural language space through multi-task contrastive learning. Then, we propose structural instruction tuning to guide LLMs in performing structure-aware reasoning over KGs, using a unified graph instruction combined with a lightweight knowledge adapter. Experimental results on two KGC tasks across four benchmark datasets demonstrate that SAT significantly outperforms state-of-the-art methods, especially in the link prediction task with improvements ranging from 8.7% to 29.8%

Text-based games (TGs) are language-based interactive environments for reinforcement learning. While language models (LMs) and knowledge graphs (KGs) are commonly used for handling large action space in TGs, it is unclear whether these techniques are necessary or overused. In this paper, we revisit the challenge of exploring the action space in TGs and propose 𝜖-admissible exploration, a minimal approach of utilizing admissible actions, for training phase. Additionally, we present a text-based actor-critic (TAC) agent that produces textual commands for game, solely from game observations, without requiring any KG or LM. Our method, on average across 10 games from Jericho, outperforms strong baselines and state-of-the-art agents that use LM and KG. Our approach highlights that a much lighter model design, with a fresh perspective on utilizing the information within the environments, suffices for an effective exploration of exponentially large action spaces.

Link prediction on knowledge graphs (KGs) has been extensively studied on binary relational KGs, wherein each fact is represented by a triple. A significant amount of important knowledge, however, is represented by hyper-relational facts where each fact is composed of a primal triple and a set of qualifiers comprising a key-value pair that allows for expressing more complicated semantics. Although some recent works have proposed to embed hyper-relational KGs, these methods fail to capture essential inference patterns of hyper-relational facts such as qualifier monotonicity, qualifier implication, and qualifier mutual exclusion, limiting their generalization capability. To unlock this, we present ShrinkE, a geometric hyper-relational KG embedding method aiming to explicitly model these patterns. ShrinkE models the primal triple as a spatial-functional transformation from the head into a relation-specific box. Each qualifier “shrinks” the box to narrow down the possible answer set and, thus, realizes qualifier monotonicity. The spatial relationships between the qualifier boxes allow for modeling core inference patterns of qualifiers such as implication and mutual exclusion. Experimental results demonstrate ShrinkE’s superiority on three benchmarks of hyper-relational KGs.

Text-based games (TGs) are exciting testbeds for developing deep reinforcement learning techniques due to their partially observed environments and large action spaces. In these games, the agent learns to explore the environment via natural language interactions with the game simulator. A fundamental challenge in TGs is the efficient exploration of the large action space when the agent has not yet acquired enough knowledge about the environment. We propose CommExpl, an exploration technique that injects external commonsense knowledge, via a pretrained language model (LM), into the agent during training when the agent is the most uncertain about its next action. Our method exhibits improvement on the collected game scores during the training in four out of nine games from Jericho. Additionally, the produced trajectory of actions exhibit lower perplexity, when tested with a pretrained LM, indicating better closeness to human language.

Abstractive summarization has made tremendous progress in recent years. In this work, we perform fine-grained human annotations to evaluate long document abstractive summarization systems (i.e., models and metrics) with the aim of implementing them to generate reliable summaries. For long document abstractive models, we show that the constant strive for state-of-the-art ROUGE results can lead us to generate more relevant summaries but not factual ones. For long document evaluation metrics, human evaluation results show that ROUGE remains the best at evaluating the relevancy of a summary. It also reveals important limitations of factuality metrics in detecting different types of factual errors and the reasons behind the effectiveness of BARTScore. We then suggest promising directions in the endeavor of developing factual consistency metrics. Finally, we release our annotated long document dataset with the hope that it can contribute to the development of metrics across a broader range of summarization settings.

This paper presents an unsupervised extractive approach to summarize scientific long documents based on the Information Bottleneck principle. Inspired by previous work which uses the Information Bottleneck principle for sentence compression, we extend it to document level summarization with two separate steps. In the first step, we use signal(s) as queries to retrieve the key content from the source document. Then, a pre-trained language model conducts further sentence search and edit to return the final extracted summaries. Importantly, our work can be flexibly extended to a multi-view framework by different signals. Automatic evaluation on three scientific document datasets verifies the effectiveness of the proposed framework. The further human evaluation suggests that the extracted summaries cover more content aspects than previous systems.

The Scholarly Document Processing (SDP) workshop is to encourage more efforts on natural language understanding of scientific task. It contains three shared tasks and we participate in the LongSumm shared task. In this paper, we describe our text summarization system, SciSummPip, inspired by SummPip (Zhao et al., 2020) that is an unsupervised text summarization system for multi-document in News domain. Our SciSummPip includes a transformer-based language model SciBERT (Beltagy et al., 2019) for contextual sentence representation, content selection with PageRank (Page et al., 1999), sentence graph construction with both deep and linguistic information, sentence graph clustering and within-graph summary generation. Our work differs from previous method in that content selection and a summary length constraint is applied to adapt to the scientific domain. The experiment results on both training dataset and blind test dataset show the effectiveness of our method, and we empirically verify the robustness of modules used in SciSummPip with BERTScore (Zhang et al., 2019a).