Searcharxiv⌕ Search

arXiv subjects

Mengru Wang

Publications and source records attributed to Mengru Wang.

At least 19 recordsLinked to original sources

Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics

Methods for controlling large language models (LLMs), including local weight fine-tuning, LoRA-based adaptation, and activation-based interventions, are often studied in isolation, obscuring their connections and making comparison difficult. In this work, we present a unified view that frames these interventions as dynamic weight updates induced by a control signal, placing them within a single conceptual framework. Building on this view, we propose a unified preference-utility analysis that separates control effects into preference, defined as the tendency toward a target concept, and utility, defined as coherent and task-valid generation, and measures both on a shared log-odds scale using polarity-paired contrastive examples. Across methods, we observe a consistent trade-off between preference and utility: stronger control increases preference while predictably reducing utility. We further explain this behavior through an activation manifold perspective, in which control shifts representations along target-concept directions to enhance preference, while utility declines primarily when interventions push representations off the model's valid-generation manifold. Finally, we introduce a new steering approach SPLIT guided by this analysis that improves preference while better preserving utility. Code is available at https://github.com/zjunlp/EasyEdit/blob/main/examples/SPLIT.md.

cs.CL↗

Winning Reviews, Losing Discovery? The Co-Evolutionary Game of AI Research and Peer Review

Autonomous research systems can produce convincing manuscripts whose scientific value is difficult to verify, while automated reviewers can reward the appearance of completeness. We examine how this gap can become a self-reinforcing threat to the research community. In a two-population evolutionary game, authors choose between acceptance-oriented and exploratory research, and reviewers choose between proxy evaluation and careful assessment. Cheaper acceptance-oriented production can increase review pressure, making proxy assessment more attractive; proxy assessment can in turn strengthen the incentive to produce reviewer-compatible work. We derive conditions under which this feedback sustains a stable regime that rewards publication success while reducing the scientific value of accepted work. An analytical basin boundary explains why the same institutional rules can support either this regime or exploratory research with careful review. Numerical illustrations show how publication counts can conceal scientific loss, how faster adjustment can reinforce existing practices, and why partial reform may fail once the cycle is established. Careful audits and reviewer rewards can reverse the feedback, while temporary support can produce lasting recovery if it moves the community across the basin boundary. The model identifies a warning mechanism rather than forecasting a particular conference: a review system can remain operational and privately rewarding while becoming less effective at supporting discovery. Its institutional implication is to reward independently validated contributions and consequential checking, rather than accelerate paper production and scoring alone.

cs.GT↗

Mechanist: AI as a Scientific Instrument for Discovering the Mechanisms of Intelligence

AI models are increasingly used in scientific discovery and human decision-making. Yet how AI models work and what risks they pose remain poorly understood. As AI development becomes faster and more automated, research on the mechanisms underlying AI remains largely manual. To bridge this gap, we introduce Mechanist, an agentic system that uses AI as a scientific instrument for the autonomous discovery of mechanisms underlying AI. To ground novel mechanism hypotheses, we construct a scientific knowledge graph of 13,000 studies on AI mechanisms, alongside a multidisciplinary database of 43 million papers spanning 26 fields. For reliable experiment execution, we curate a library of 32 foundational methods for mechanism analysis. Compared with Claude Code and existing AI-scientist systems, Mechanist generates higher-quality mechanism hypotheses and executes experiments more reliably. Across four case studies, Mechanist autonomously discovers new model behaviors and their underlying mechanisms, and translates these discoveries into mechanism-guided interventions and interdisciplinary design. Specifically, Mechanist first uncovers a counterintuitive safety risk in scientific laboratories, showing that unsafe traits can transfer to fine-tuned student models through apparently safe training data and emerge across modalities. Mechanist then develops a mechanism theory of belief, revealing how models represent world knowledge, form beliefs, infer the beliefs of others, and how these mechanisms emerge during pretraining. Building on this theory, Mechanist develops targeted interventions that improve model performance across diverse scenarios. Finally, Mechanist can also advance interdisciplinary discovery through mechanistic design, providing an alternative to the computationally intensive generate-and-rerank paradigm.

cs.AI↗

When Should Models Change Their Minds? Contextual Belief Management in Large Language Models

Long-horizon interactions require language models to manage accumulating information: when to update their state, when to preserve their state, and what to ignore. We study this challenge as Contextual Belief Management (CBM): maintaining a predicted belief state aligned with formal evidence while isolating task-irrelevant noise. To make CBM measurable, we introduce BeliefTrack, a closed-world benchmark spanning Rule Discovery and Circuit Diagnosis, where a finite belief space and symbolic verifiers enable exact turn-level evaluation. BeliefTrack diagnoses three failures: Failed Stay, Failed Update, and Failed Isolation. Across multiple LLMs, vanilla models exhibit severe CBM failures, while explicit belief-tracking prompts provide limited gains. In contrast, reinforcement learning with belief-state rewards reduces failure rates by 70.9% on average. Further probing reveals latent belief-state dynamics behind these failures, and representation-level steering reduces failure rates by 46.1% across two tasks (Code is available at https://github.com/zjunlp/CBM).

cs.AI↗

SciAtlas: A Computable Atlas of Science for Knowledge-Grounded AI Research

Artificial intelligence is rapidly entering the core workflows of scientific research. Yet reliable scientific reasoning requires access to accumulated scientific knowledge with sufficient breadth, depth, and standardization. Current AI scientists typically assemble scientific knowledge through workflow- and discipline-specific pipelines, which provide incomplete coverage, leave relations implicit, and make knowledge acquisition pathways fragmented. Here we present SciAtlas, a shared, machine-actionable cross-disciplinary scholarly knowledge infrastructure that integrates evidential, conceptual, disciplinary, expertise, and normative layers under a shared schema. SciAtlas further achieves a unified neuro-symbolic retrieval mechanism that grounds heterogeneous research objects, propagates relevance across the scholarly topology, and projects the resulting relevance field into the context required by each scientific workflow. Across three representative workflows, SciAtlas broadens trajectory reconstruction by recovering overlooked research branches, deepens opportunity discovery by uncovering underexplored bottlenecks and cross-domain insights, and strengthens innovation assessment by integrating evidence, expertise, and evaluation signals. Across three representative workflows, SciAtlas broadens trajectory reconstruction by recovering overlooked stages and branches, deepens opportunity discovery by uncovering underexplored bottlenecks and cross-domain connections, and standardizes innovation assessment by integrating evidence, expertise and evaluation signals. Extensive evaluations validate the foundational capabilities underpinning it as reusable knowledge infrastructure for knowledge-intensive scientific research.

cs.AI↗

OceanGym: A Benchmark Environment for Underwater Embodied Agents

We introduce OceanGym, the first comprehensive benchmark for ocean underwater embodied agents, designed to advance AI in one of the most demanding real-world environments. Unlike terrestrial or aerial domains, underwater settings present extreme perceptual and decision-making challenges, including low visibility, dynamic ocean currents, making effective agent deployment exceptionally difficult. OceanGym encompasses eight realistic task domains and a unified agent framework driven by Multi-modal Large Language Models (MLLMs), which integrates perception, memory, and sequential decision-making. Agents are required to comprehend optical and sonar data, autonomously explore complex environments, and accomplish long-horizon objectives under these harsh conditions. Extensive experiments reveal substantial gaps between state-of-the-art MLLM-driven agents and human experts, highlighting the persistent difficulty of perception, planning, and adaptability in ocean underwater environments. By providing a high-fidelity, rigorously designed platform, OceanGym establishes a testbed for developing robust embodied AI and transferring these capabilities to real-world autonomous ocean underwater vehicles, marking a decisive step toward intelligent agents capable of operating in one of Earth's last unexplored frontiers. The code and data are available at https://github.com/OceanGPT/OceanGym.

cs.CL↗

SkillNet: Create, Evaluate, and Connect AI Skills

Current AI agents can flexibly invoke tools and execute complex tasks, yet their long-term advancement is hindered by the lack of systematic accumulation and transfer of skills. Without a unified mechanism for skill consolidation, agents frequently ``reinvent the wheel'', rediscovering solutions in isolated contexts without leveraging prior strategies. To address this challenge, we introduce SkillNet, an open infrastructure for creating, evaluating, and organizing AI skills at scale. SkillNet structures skills within a unified ontology that supports creating skills from heterogeneous sources, establishing rich relational connections, and performing multi-dimensional evaluation across Safety, Completeness, Executability, Maintainability, and Cost-awareness. Our infrastructure integrates a repository of over 600,000 skills, an interactive platform, and a versatile Python toolkit. Experiments on ALFWorld, WebShop, and ScienceWorld show 40% higher average rewards and 30% fewer execution steps across multiple backbone models. Furthermore, SkillNet-Gym benchmarks skill retrieval, utilization, and composition, while SkillNet-Fabric enables task-specific skill routing through lightweight Wikis. By formalizing skills as evolving, composable assets, SkillNet provides a robust foundation for agents to move from transient experience to durable mastery.

cs.AI↗

MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use

Memory has become a key component of large language models, enabling them to retain information and learn from long-term interactions. However, existing memory benchmarks mainly evaluate whether information is correctly extracted, stored, and retrieved, while largely overlooking how retrieved memories reshape model reasoning and affect performance on the current task. We identify memory-induced cognitive traps: even faithfully recorded and semantically relevant memories can distort model reasoning or beliefs and degrade current task performance. To systematically evaluate these failure modes, we introduce MemTrapBench, which covers two forms of cognitive traps: Reasoning Fixation and Belief Distortion. Experiments across two model families and five representative memory frameworks show that MemTrapBench is challenging: all evaluated memory strategies underperform the no-memory setting, with even the strongest methods suffering drops of more than 10%. To mitigate these cognitive traps, we propose AdaptiveMem, a simple yet effective inference-time method that instructs LLMs to avoid memory traps. AdaptiveMem mitigates cognitive traps on MemTrapBench while preserving or improving performance on standard memory benchmarks across diverse memory frameworks.

cs.AI↗

MobileMem: Learning from a Year of Mobile Experiences

The next generation of AI agents is increasingly moving beyond systems that answer isolated questions toward persistent personal assistants that can understand, remember, and continuously learn from users' experiences. Such assistants require long-term memory to accumulate and leverage user-specific experiences over time, yet existing benchmarks remain inadequate for realistic mobile settings, where experiences are heterogeneous, multimodal, evolving, and deeply personal. We introduce MobileMem, a benchmark and framework for studying on-device long-term memory, grounded in a year-scale collection of mobile experiences. MobileMem employs a knowledge-grounded synthesis pipeline to construct coherent and temporally consistent long-horizon trajectories from user-app sessions. It provides complementary text and multimodal settings covering multi-hop and temporal reasoning, knowledge updating, and implicit preference inference. Specifically, MobileMem enables agents to remember the past, understand the present, and adapt to the future. By modeling experiences rather than isolated facts, MobileMem moves memory beyond information retrieval toward experiential intelligence for continuous personal learning.

cs.AI↗

IEMAS: An Incentive-Efficiency Routing Framework for Open Agentic Web Ecosystems

The transition to open, distributed Multi-Agent Systems (MAS) promises scalable intelligence but introduces a non-trivial tension: maximizing global efficiency requires cooperative, resource-aware scheduling, yet autonomous agents may be self-interested and cannot be managed by a centralized controller. Prior approaches fall short in two key areas: they typically focus on single-query routing, neglecting long-term resource reuse (e.g., KV-caching) and the complexities of system-level many-to-many matching; furthermore, they rely on generic incentive mechanisms that ignore the distinct characteristics of LLM inference. To bridge this gap, we propose IEMAS (Incentive-Efficiency Mechanism for Multi-Agent Systems), a distributed framework that aligns economic incentives with system performance. IEMAS integrates a probabilistic predictive model to estimate Quality of Service (QoS) under uncertainty, which feeds into a VCG-based bipartite matching mechanism. This design guarantees truthful capability reporting and social optimality while explicitly leveraging KV cache-affinity to minimize computational redundancy. We implement IEMAS on top of vLLM and evaluate it via extensive simulations. Results demonstrate that our incentive-efficiency co-design reducing average service cost by 35% and end-to-end latency by up to 2.9 compared to baselines.

cs.NI↗

LightMem: Lightweight and Efficient Memory-Augmented Generation

Despite their remarkable capabilities, Large Language Models (LLMs) struggle to effectively leverage historical interaction information in dynamic and complex environments. Memory systems enable LLMs to move beyond stateless interactions by introducing persistent information storage, retrieval, and utilization mechanisms. However, existing memory systems often introduce substantial time and computational overhead. To this end, we introduce a new memory system called LightMem, which strikes a balance between the performance and efficiency of memory systems. Inspired by the Atkinson-Shiffrin model of human memory, LightMem organizes memory into three complementary stages. First, cognition-inspired sensory memory rapidly filters irrelevant information through lightweight compression and groups information according to their topics. Next, topic-aware short-term memory consolidates these topic-based groups, organizing and summarizing content for more structured access. Finally, long-term memory with sleep-time update employs an offline procedure that decouples consolidation from online inference. On LongMemEval and LoCoMo, using GPT and Qwen backbones, LightMem consistently surpasses strong baselines, improving QA accuracy by up to 7.7% / 29.3%, reducing total token usage by up to 38x / 20.9x and API calls by up to 30x / 55.5x, while purely online test-time costs are even lower, achieving up to 106x / 117x token reduction and 159x / 310x fewer API calls. The code is available at https://github.com/zjunlp/LightMem.

cs.CL↗

From Data to Behavior: Predicting Unintended Model Behaviors Before Training

Large Language Models (LLMs) can acquire unintended biases from seemingly benign training data even without explicit cues or malicious content. Existing methods struggle to detect such risks before fine-tuning, making post hoc evaluation costly and inefficient. To address this challenge, we introduce Data2Behavior, a new task for predicting unintended model behaviors prior to training. We also propose Manipulating Data Features (MDF), a lightweight approach that summarizes candidate data through their mean representations and injects them into the forward pass of a base model, allowing latent statistical signals in the data to shape model activations and reveal potential biases and safety risks without updating any parameters. MDF achieves reliable prediction while consuming only about 20% of the GPU resources required for fine-tuning. Experiments on Qwen3-14B, Qwen2.5-32B-Instruct, and Gemma-3-12b-it confirm that MDF can anticipate unintended behaviors and provide insight into pre-training vulnerabilities.

cs.LG↗

LookAhead Tuning: Safer Language Models via Partial Answer Previews

Fine-tuning enables large language models (LLMs) to adapt to specific domains, but often compromises their previously established safety alignment. To mitigate the degradation of model safety during fine-tuning, we introduce LookAhead Tuning, a lightweight and effective data-driven approach that preserves safety during fine-tuning. The method introduces two simple strategies that modify training data by previewing partial answer prefixes, thereby minimizing perturbations to the model's initial token distributions and maintaining its built-in safety mechanisms. Comprehensive experiments demonstrate that LookAhead Tuning effectively maintains model safety without sacrificing robust performance on downstream tasks. Our findings position LookAhead Tuning as a reliable and efficient solution for the safe and effective adaptation of LLMs.

cs.CL↗

Too Good to be Bad: On the Failure of LLMs to Role-Play Villains

Large Language Models (LLMs) are increasingly tasked with creative generation, including the simulation of fictional characters. However, their ability to portray non-prosocial, antagonistic personas remains largely unexamined. We hypothesize that the safety alignment of modern LLMs creates a fundamental conflict with the task of authentically role-playing morally ambiguous or villainous characters. To investigate this, we introduce the Moral RolePlay benchmark, a new dataset featuring a four-level moral alignment scale and a balanced test set for rigorous evaluation. We task state-of-the-art LLMs with role-playing characters from moral paragons to pure villains. Our large-scale evaluation reveals a consistent, monotonic decline in role-playing fidelity as character morality decreases. We find that models struggle most with traits directly antithetical to safety principles, such as ``Deceitful'' and ``Manipulative'', often substituting nuanced malevolence with superficial aggression. Furthermore, we demonstrate that general chatbot proficiency is a poor predictor of villain role-playing ability, with highly safety-aligned models performing particularly poorly. Our work provides the first systematic evidence of this critical limitation, highlighting a key tension between model safety and creative fidelity. Our benchmark and findings pave the way for developing more nuanced, context-aware alignment methods.

cs.CL↗

Social Welfare Function Leaderboard: When LLM Agents Allocate Social Welfare

Large language models (LLMs) are increasingly entrusted with high-stakes decisions that affect human welfare. However, the principles and values that guide these models when distributing scarce societal resources remain largely unexamined. To address this, we introduce the Social Welfare Function (SWF) Benchmark, a dynamic simulation environment where an LLM acts as a sovereign allocator, distributing tasks to a heterogeneous community of recipients. The benchmark is designed to create a persistent trade-off between maximizing collective efficiency (measured by Return on Investment) and ensuring distributive fairness (measured by the Gini coefficient). We evaluate 20 state-of-the-art LLMs and present the first leaderboard for social welfare allocation. Our findings reveal three key insights: (i) A model's general conversational ability, as measured by popular leaderboards, is a poor predictor of its allocation skill. (ii) Most LLMs exhibit a strong default utilitarian orientation, prioritizing group productivity at the expense of severe inequality. (iii) Allocation strategies are highly vulnerable, easily perturbed by output-length constraints and social-influence framing. These results highlight the risks of deploying current LLMs as societal decision-makers and underscore the need for specialized benchmarks and targeted alignment for AI governance.

cs.CL↗

The Hunger Game Debate: On the Emergence of Over-Competition in Multi-Agent Systems

LLM-based multi-agent systems demonstrate great potential for tackling complex problems, but how competition shapes their behavior remains underexplored. This paper investigates the over-competition in multi-agent debate, where agents under extreme pressure exhibit unreliable, harmful behaviors that undermine both collaboration and task performance. To study this phenomenon, we propose HATE, the Hunger Game Debate, a novel experimental framework that simulates debates under a zero-sum competition arena. Our experiments, conducted across a range of LLMs and tasks, reveal that competitive pressure significantly stimulates over-competition behaviors and degrades task performance, causing discussions to derail. We further explore the impact of environmental feedback by adding variants of judges, indicating that objective, task-focused feedback effectively mitigates the over-competition behaviors. We also probe the post-hoc kindness of LLMs and form a leaderboard to characterize top LLMs, providing insights for understanding and governing the emergent social dynamics of AI community.

cs.CL↗

Automating Steering for Safe Multimodal Large Language Models

Recent progress in Multimodal Large Language Models (MLLMs) has unlocked powerful cross-modal reasoning abilities, but also raised new safety concerns, particularly when faced with adversarial multimodal inputs. To improve the safety of MLLMs during inference, we introduce a modular and adaptive inference-time intervention technology, AutoSteer, without requiring any fine-tuning of the underlying model. AutoSteer incorporates three core components: (1) a novel Safety Awareness Score (SAS) that automatically identifies the most safety-relevant distinctions among the model's internal layers; (2) an adaptive safety prober trained to estimate the likelihood of toxic outputs from intermediate representations; and (3) a lightweight Refusal Head that selectively intervenes to modulate generation when safety risks are detected. Experiments on LLaVA-OV and Chameleon across diverse safety-critical benchmarks demonstrate that AutoSteer significantly reduces the Attack Success Rate (ASR) for textual, visual, and cross-modal threats, while maintaining general abilities. These findings position AutoSteer as a practical, interpretable, and effective framework for safer deployment of multimodal AI systems.

cs.CL↗

EasyEdit2: An Easy-to-use Steering Framework for Editing Large Language Models

In this paper, we introduce EasyEdit2, a framework designed to enable plug-and-play adjustability for controlling Large Language Model (LLM) behaviors. EasyEdit2 supports a wide range of test-time interventions, including safety, sentiment, personality, reasoning patterns, factuality, and language features. Unlike its predecessor, EasyEdit2 features a new architecture specifically designed for seamless model steering. It comprises key modules such as the steering vector generator and the steering vector applier, which enable automatic generation and application of steering vectors to influence the model's behavior without modifying its parameters. One of the main advantages of EasyEdit2 is its ease of use-users do not need extensive technical knowledge. With just a single example, they can effectively guide and adjust the model's responses, making precise control both accessible and efficient. Empirically, we report model steering performance across different LLMs, demonstrating the effectiveness of these techniques. We have released the source code on GitHub at https://github.com/zjunlp/EasyEdit along with a demonstration notebook. In addition, we provide a demo video at https://www.youtube.com/watch?v=AkfoiPfp5rQ for a quick introduction.

cs.CL↗