Computation and Language
MoG: Mixture of Experts for Graph-based Retrieval-Augmented Generation
Zheng Yuan, Chuang Zhou, Linhao Luo et al. (+4 more)
Retrieval-augmented generation is intensively studied to ground large
language models on external evidence. However, retrieving from a unified knowledge base could inevitably introduce irrelevant information that may mislead generation for complex
reasoning. Inspired by the conditional computation of mixture of experts (MoE), where a router sparsely selects specialized experts alongside shared ones for each input, we propose Mixture of experts for Graph-based
Retrieval-Augmented Generation, i.e., MoG. It organizes knowledge into two core components: (i) diverse, always-accessible hub graphs that encode semantically and structurally central knowledge and provide contextual clues for expert activation, and (ii) sparsely activated expert graphs that contain domain-specific evidence. MoG first accesses hub graphs to identify general evidence and derive contextual clues. Then, a
topology-aware router dynamically activates a limited set of expert graphs conditioned on the query, thereby confining
retrieval to a focused evidence subspace. Extensive experiments on challenging
benchmarks show that MoG consistently outperforms strong baselines, with over 20% relative improvement on MuSiQue. Our code is available in https://github.com/DEEP-PolyU/MoG.
Read the full abstract ▾
All authors Zheng Yuan, Chuang Zhou, Linhao Luo, Siyu An, Di Yin, Xing Sun, Xiao Huang
Computation and Language · Cross-listed from cs.CY
Traceable by Design: An LLM Pipeline and Dashboard for EU Regulatory Consultation Analysis
Thales Bertaglia, Haoyang Gui, Catalina Goanta et al. (+1 more)
Public consultations generate large volumes of data in the form of stakeholder submissions that are practically unfeasible to analyse manually. We present an end-to-end LLM-based pipeline and interactive dashboard for structured topic extraction from regulatory consultation submissions, demonstrated on the European Commission's Digital
Fairness Act (DFA) public call for evidence as a case study. The system processes raw PDF attachments and web-form responses, extracts topic annotations, and grounds every extraction in a verbatim quote from the source text. Applied to 4,322 DFA submissions, the pipeline produced 15,368 topic annotations supported by 20,951 verbatim evidence quotes. Three principles govern the proposed design: verbatim grounding, full traceability, and transparency by design. The dashboard exposes the full extraction
dataset through five analytical views, from
dataset-level topic overviews to individual paragraph drill-downs, with every result traceable to its source. Beyond the predefined DFA topic categories, the pipeline generated certain stakeholder concerns, such as Age Verification, Payment Processor Censorship, and Digital Ownership, that a fixed-taxonomy approach would have missed. The pipeline is domain-generic; adapting it to a new consultation requires only a
prompt update and a new
dataset. A live demo is available at https://dfa-dashboard.thalesbertaglia.com/. The code and processed data are publicly available at https://github.com/thalesbertaglia/dfa-dashboard.
Read the full abstract ▾
All authors Thales Bertaglia, Haoyang Gui, Catalina Goanta, Gerasimos Spanakis
Computation and Language · Cross-listed from cs.CV
Generating Reports or Repeating Templates? Measuring and Mitigating Template Collapse in 3D CT Report Generation
Tom Maye-Lasserre, Yitong Li, Bailiang Jian et al. (+3 more)
Modern 3D medical
vision-
language models (VLMs) can generate fluent radiology-style text while exhibit critically low pathology detection and output diversity, collapsing to generic templates that under-report rare yet critical findings. We identify this failure mode as Template Collapse. This failure stems from the unique constraints of 3D medical imaging, e.g., limited data, severe label imbalance, and weak signals from volumetric encoders. Under these constraints, text-generation objectives encourage shortcut learning and fluent but weakly grounded reports. We systematically diagnose the Template Collapse through clinical fidelity, output diversity, normal-template bias, and rare-finding survival. To mitigate it, we propose CLarGen, a decoupled framework that separates what to say (clinical detection) from how to say it (
language synthesis). CLarGen uses (i) a Latent Query
Transformer for multi-label pathology detection, (ii) pathology-guided
retrieval for clinically matched exemplars, and (iii) a medical
language model to synthesize the final report from detected findings and retrieved context. Across state-of-the-art 3D CT report generation baselines, CLarGen mitigates Template Collapse and substantially improves clinical accuracy (macro-F1 0.487 vs. 0.189; CRG 0.472 vs. 0.368) while maintaining fluent reporting. Our results suggest that explicit, measurable clinical grounding is essential for template-collapse-resistant 3D CT report generation. Code is available at https://github.com/ai-med/CLarGen.
Read the full abstract ▾
All authors Tom Maye-Lasserre, Yitong Li, Bailiang Jian, Morteza Ghahremani, Benedikt Wiestler, Christian Wachinger
Computation and Language
Cognitive Fatigue in Autoregressive Transformers: Formalization and Measurement
Riju Marwah, Ritvik Garimella, Vishal Pallagani et al. (+3 more)
Autoregressive
language models frequently degrade during long-horizon generation, producing repetitive text, losing instruction adherence, and exhibiting unstable
entropy. Despite the prevalence of these failures, practitioners lack online diagnostics to detect them in real-time as they occur. We formalize this degradation as cognitive fatigue, a measurable generation-time state characterized by decay in
attention to the original
prompt, representational drift, and
entropy miscalibration. We introduce the Fatigue Index (FI), a lightweight, model-agnostic diagnostic that aggregates these three signals under explicit axioms (monotonicity, boundedness, interpretability) enabling reliable runtime monitoring. Across nine models (1B-13B parameters), FI trajectories exhibit structured temporal dynamics, predict task degradation (AUROC = 0.95) and repetition (Spearman rho = 0.94), and reveal non-monotonic scaling behavior: instruction-tuned models below 3B exhibit faster collapse than base models, with this trend reversing at 7B. Stress analyses further show that FI onset accelerates under longer contexts, middle-positioned evidence, and reduced numerical precision. These results establish cognitive fatigue as a coherent and measurable phenomenon, and position FI as a principled tool for runtime reliability monitoring in production LLM systems.
Read the full abstract ▾
All authors Riju Marwah, Ritvik Garimella, Vishal Pallagani, Atishay Jain, Michael Stewart, Amit Sheth
Computation and Language · Cross-listed from cs.IR
Reading Between the Citations: A Typed Claim Network for Scientific Literature
Ning Ding, Sergio J. Rodríguez Méndez and Pouya G. Omran
Knowledge graphs over corpora of inter-referencing documents - scholarly papers, legal opinions, policy briefs - encode the
topology of reference but not its stance. The standard representation collapses a rich evaluative relation into an untyped edge, losing the very content that supports community-level queries about how one document is received by another. We propose the claim network: a representational pattern in which each cross-document reference is reified as a typed claim, carrying source, target, claim text, and a four-class stance label grounded in the citation-intent literature. We give a construction pipeline applicable to any corpus of scholarly inter-referencing documents and instantiate it on a corpus of 127 papers in 3D point cloud semantic
segmentation, producing a network of 8,260 typed claims. Three downstream task families demonstrate what the network enables:
retrieval signal augmentation, aggregated-stance summarisation, and topological analytics. Head-to-head evaluation against standard
Retrieval-Augmented Generation (RAG) baselines shows that the gain over flat
retrieval is the gain from the right intermediate representation rather than the wrong one.
Read the full abstract ▾
Computation and Language · Cross-listed from eess.AS
Jun-Hak Yun, Seung-Bin Kim and Seong-Whan Lee
Recent advancements in text-guided audio generation have yielded promising results in diverse domains, including sound effects,
speech, and music. However, jointly generating
speech with environmental audio remains challenging due to the inherent disparities in their acoustic patterns and temporal dynamics. We propose ImmersiveTTS, an environment-aware text-to-
speech (TTS) model that generates natural
speech seamlessly integrated within environmental contexts by explicitly modeling cross-modal interactions. Our model builds on a
multimodal diffusion transformer and fuses transcript-aligned
speech latent with text-conditioned environmental context via joint
attention. To enhance semantic consistency, we introduce a domain-specific representation
alignment objective tailored to environment-aware TTS, leveraging complementary self-supervised representations from
speech and audio encoders. Experimental results show that ImmersiveTTS achieves higher naturalness, intelligibility, and audio fidelity than existing approaches across objective metrics and human listening tests.
Read the full abstract ▾
Computation and Language
EvoGens: A Population-Based Heuristic Search Framework for Scientific Idea Generation
Xu Li, Hanzhe Tu, Xinyi Li et al. (+3 more)
Generating novel research ideas is fundamental to scientific progress. While Large
Language Models (LLMs) show promise in assisting this process, existing approaches often exhibit semantic convergence, resulting in limited diversity and novelty. To address this, we introduce EvoGens, an
evolution-inspired framework that recasts scientific idea generation as an evolutionary search over a population of ideas. EvoGens iteratively applies rank-based mutation with differentiated
retrieval planning to incorporate external knowledge, and semantic-aware crossover to fuse complementary concepts for conceptual reorganization. A lightweight evaluation signal guides the selection process, encouraging sustained exploration while mitigating premature convergence. Extensive experiments demonstrate that EvoGens substantially enhances exploration capabilities compared to state-of-the-art baselines. Specifically, it improves the Novelty from 0.1 to 0.4 and the Diversity from 0.24 to 0.55, while maintaining comparable idea quality under the current automatic evaluation protocol. These findings suggest that evolutionary mechanisms can serve as a useful framework for exploration-oriented research ideation, especially for broadening the novelty and diversity of candidate ideas under a shared automatic evaluation setting.
Read the full abstract ▾
All authors Xu Li, Hanzhe Tu, Xinyi Li, Kuncheng Zhao, Xun Han, Zhonghui Liu
Computation and Language
Extending AI for Research to the Humanities: A Multi-Agent Framework for Evidence-Grounded Scholarship
Yating Pan, Jiajun Zhang, Jun Wang et al. (+1 more)
LLM-based research
agents have advanced rapidly in science and engineering, where research is organized around executable experiments, code, and quantitative signals. Humanities scholarship, however, requires interpretive, evidence-grounded arguments over primary sources, whose value rests on faithful quotation, verifiable provenance, and deep interpretation. Existing research
agents use general planning, tool use, and reflection, leaving these scholarly operations implicit. To address this gap, we introduce SPIRE (Scholarly-Primitives-Inspired Research Engine), a multi-
agent framework that realizes Scholarly Primitives, a typology of basic humanities scholarship practices, as cooperating roles over a multi-scale close-reading substrate of passages, intra-context graph communities, and cross-context semantic clusters. On a
benchmark of peer-reviewed papers in classical Chinese and Greco-Roman Latin scholarship, SPIRE recovers substantially more cited primary-source evidence and receives higher blind human and LLM ratings on four scholarly quality dimensions than LLM, RAG, and generic agentic baselines.
Retrieval-tier and
agent ablations, with
retrieval-volume and answer-length controls, trace its advantage to targeted
retrieval, structured evidence selection, and claim-evidence binding. Code and
benchmark data are released at https://github.com/YatingPan/SPIRE.
Read the full abstract ▾
All authors Yating Pan, Jiajun Zhang, Jun Wang, Qi Su
Computation and Language
Do Large Language Models Encode Institutional Experience? Evidence from Cross-Linguistic Moral Reasoning Under Ambiguity
Nattavudh Powdthavee
Large
language models (LLMs) exhibit systematic differences in moral
reasoning across
languages, yet the source of this variation remains unclear. We test the hypothesis that
languages encode aspects of the institutional environments in which they are spoken, allowing LLMs to inherit institution-specific moral priors through training. Across nine
languages spanning a broad gradient of institutional quality, six frontier LLMs, and two preregistered studies, we examine moral dilemmas whose acceptability depends on institutional functioning. In Study 1, explicit institutional framing produced uniformly null results: cross-linguistic moral divergence did not increase in institutionally contingent scenarios, nor did it track institutional differences between
language communities. In Study 2, we introduced institutionally ambiguous scenarios in which institutional stakes were present but not explicitly stated. Under these conditions, cross-linguistic moral divergence increased relative to institutionally inert controls and, with one theoretically informative exception, was associated with real-world institutional differences between
language communities. Explicit framing again attenuated these effects. These findings suggest that institutional experience may leave detectable traces in
language that shape LLM moral
reasoning, while also indicating that explicit institutional cues can suppress the expression of those differences.
Read the full abstract ▾
Computation and Language
MineExplorer: Evaluating Open-World Exploration of MLLM Agents in Minecraft
Tianjie Ju, Yueqing Sun, Zheng Wu et al. (+7 more)
Multimodal large
language models (MLLMs) have shown strong capabilities in perception,
reasoning, and action generation. However, their ability to sustain exploration in dynamic open worlds remains unclear. Existing
embodied and game-based
benchmarks often compress interaction into short-horizon tasks or entangle success with domain-specific game mechanics. In this paper, we introduce MineExplorer
benchmark for evaluating open-world exploration capabilities of MLLM
agents in Minecraft. We first filter atomic tasks whose solutions rely heavily on Minecraft-specific knowledge to better reflect general open-world
reasoning. Then we organize the
benchmark around a ReAct-style capability formulation and compose atomic tasks into implicit multi-hop tasks. To further construct reliable instances, MineExplorer uses a multi-
agent synthesis workflow that jointly designs task graphs, sandbox scenes, and rule-based milestone evaluators. Human evaluation shows that the multi-
agent synthesis workflow produces significantly more reliable instances than a single-
agent baseline. Experiments with advanced MLLM
agents show that open-world exploration remains challenging, as strong models can handle many single-hop tasks but degrade sharply when hidden prerequisites must be coordinated over longer trajectories. Further analysis finds that task difficulty tracks
agent completion, and larger models or thinking modes do not consistently translate into better performance. Code and
dataset are available at https://github.com/meituan-longcat/MineExplorer.
Read the full abstract ▾
All authors Tianjie Ju, Yueqing Sun, Zheng Wu, Wei Zhang, Yaqi Huo, Xi Su, Qi Gu, Xunliang Cai, Gongshen Liu, Zhuosheng Zhang
Computation and Language · Cross-listed from cs.HC
TUX: Measuring Human–AI Tacit Understanding
Yueshen Li, Hanyi Min, Vedant Das Swain et al. (+1 more)
As large
language models (LLMs) increasingly act as collaborative partners, human–AI
alignment is often evaluated through explicit task success, accuracy, or reward optimization. Yet many collaborative settings depend on tacit understanding: whether an
agent can align with a human's evaluative stance or representational priors without clear objectives, communication, or feedback. To study this capacity, we develop a spectrum-placement task inspired by the social party game Wavelength, in which humans and
agents independently place concepts along subjective spectra. We operationalize the Tacit Understanding Index (TUX) as a pairwise behavioral measure of similarity between human and
agent judgments, and evaluate it with 241 human participants and 200 profile-conditioned LLM
agents across four models. We find that nearest human–
agent pairs in trait space achieve significantly higher TUX, suggesting that tacit
alignment is associated with person-level characteristics rather than reflecting only random similarity. Regression analyses show that TUX becomes more explainable as predictor sets become richer, with individual traits, decision-making styles, and confidence improving over aggregate trait-distance baselines. These findings suggest that TUX provides a measurable behavioral signal of human–LLM tacit understanding, while revealing the limits of profile-based conditioning for capturing deeper representational
alignment.
Read the full abstract ▾
All authors Yueshen Li, Hanyi Min, Vedant Das Swain, Koustuv Saha
Computation and Language
EMBGuard: Constructing Hazard-Aware Guardrails for Safe Planning in Embodied Agents
Dongwook Choi, Taeyoon Kwon, Bogyung Jeong et al. (+6 more)
MLLM-powered
embodied agents deployed in real-world environments encounter physical hazards. However, existing approaches lack explicit mechanisms for identifying hazards and
reasoning about action-conditioned risks, leading
agents to either miss risky interactions or over-identify risks. To address this, we propose EMBGuard, the first MLLM-based safety guardrail for
embodied agents designed to decouple physical risk
reasoning from
agent policy. By evaluating a (visual observation, action) pair, EMBGuard identifies hazardous configurations and provides natural
language explanations of potential risks. Alongside EMBGuard, we contribute EMBHazard, a training
dataset of 15.1K action-conditioned pairs, and EMBGuardTest, a
benchmark of 329 manually curated real-world scenarios spanning seven physical risk categories. Through compositional variation of hazards and actions, we generate diverse risky and benign scenarios that
agents may encounter during planning. Despite its compact size (2B, 4B), EMBGuard achieves performance competitive with proprietary MLLMs (e.g., GPT-5.1, Gemini-2.5-Pro) while significantly reducing the false-positive rates that hinder real-time deployment. We make the code, data, and models publicly available at https://github.com/dongwxxkchoi/EMBGuard
Read the full abstract ▾
All authors Dongwook Choi, Taeyoon Kwon, Bogyung Jeong, Minju Kim, Yeonjun Hwang, Hyojun Kim, Byungchul Kim, Young Kyun Jang, Jinyoung Yeo
Computation and Language
Prompt PerturbAItions: Mechanisms of LLM Response Inconsistencies
Soorya Ram Shimgekar, Agam Goyal, Amruta Parulekar et al. (+6 more)
Large
language models (LLMs) are increasingly deployed in conversational settings where user phrasing ranges from polite to casual to toxic, yet less is known about whether such lexical variation in otherwise semantically equivalent
prompts can degrade factual reliability. We study how lexical and tone-based
prompt perturbations affect the factual reliability of LLMs. Using controlled
prompt variations spanning polite, random, and increasingly toxic levels, we evaluate seven LLMs on ARC-Easy, GSM8K, and MMLU. We find that lexical perturbations reduce factual accuracy and increase predictive uncertainty in most models. Interestingly, increasing toxicity results in decreasing accuracy. To examine whether these answer inconsistencies correspond to internal changes, we conduct attribution-graph analyses of model activations and influences. We find that perturbations amplify perturbation-sensitive variant nodes relative to core
reasoning nodes. We also observed a monotonic trend with increasing toxicity score and increasing activation and influence in Variant nodes. These findings position
prompt wording and tone as a critical dimension of LLM reliability and provide behavioral and mechanistic evidence that surface-level lexical variation can alter factual outputs and internal computation.
Read the full abstract ▾
All authors Soorya Ram Shimgekar, Agam Goyal, Amruta Parulekar, Joshua Chen, Yian Wang, Navin Kumar, Hari Sundaram, Eshwar Chandrasekharan, Koustuv Saha
Computation and Language · Cross-listed from cs.CV
Attend to Evidence: Evidence-Anchored Spatial Attention Supervision for Multimodal RLVR
Ruina Hu, Chen Wang, Lai Wei et al. (+5 more)
Reinforcement learning with verifiable rewards (RLVR) improves
vision-
language models (VLMs) by optimizing outcome rewards derived from final answers. However, such outcome-only rewards do not tell the model which image regions justify an answer. For questions that require visual grounding, these rewards cannot distinguish responses supported by relevant visual evidence from those produced by
language-prior shortcuts or lucky guesses. We introduce EASE (Evidence-Anchored Spatial
Attention), which augments
multimodal RLVR with visual-evidence process supervision. EASE converts annotated evidence regions into a smoothed visual-
token target and uses it to guide response-to-image
attention during RL training, but only on high-reward trajectories. The annotations are used solely as privileged training labels, while inference requires only the original image and question. Across Qwen2.5-VL-7B, Qwen3-VL-4B, and Qwen3-VL-8B, EASE raises average scores over DAPO by 2.5 to 3.1 points on perception,
hallucination, visual math, and
multimodal reasoning benchmarks. Diagnostics and ablations show that EASE better aligns visual
attention with annotated evidence regions.
Read the full abstract ▾
All authors Ruina Hu, Chen Wang, Lai Wei, Jionghao Bai, Bin Yu, Weiran Huang, Kai Wang, Yue Wang
Computation and Language · Cross-listed from cs.SE
BlueFin: Benchmarking LLM Agents on Financial Spreadsheets
Srivatsa Kundurthy, Clara Na, Colton Moraine et al. (+6 more)
We present BlueFin, a
benchmark that tasks large
language model (LLM)
agents with synthesis, manipulation, and comprehension tasks over spreadsheet workbooks in the professional finance domain. Though estimates of the global population of paying users of spreadsheet software range in the hundreds of millions – an order of magnitude more than the estimated global population of professional developers – comparatively fewer resources have been devoted to exploring and expanding LLM capabilities in the spreadsheet domain, with fewer still dedicated to mirroring real occupational tasks encountered by those in professional finance roles. In response, we curate a set of 131 challenging, complex tasks with real-world relevance in the domain, containing 3,225 granular rubric criteria; notably, our rubric criteria and LM judge evaluations are validated by a team of expert human annotators, resulting in high-quality, granular evaluations of complex tasks that are difficult to verify programmatically but can be reliably evaluated by an LM judge
agent. Our judge achieves parity with expert consensus ($\alpha=0.826$) with a macro-F1 score of 0.839. Frontier LLMs demonstrate poor performance on the challenging
benchmark, with the strongest LLMs achieving less than 50% average scores across tasks – models exhibit particular weaknesses in dynamic correctness. Our contributions include a
dataset of examples across three categories of spreadsheet tasks, an open source harness and agentic evaluation framework, and a characterization of existing frontier models' performance on our
benchmark.
Read the full abstract ▾
All authors Srivatsa Kundurthy, Clara Na, Colton Moraine, Anoushka Mohta, Case Winter, George Fang, John Ling, Emma Strubell, Zach Kirshner
Computation and Language · Cross-listed from cs.AI
UniScale: Adaptive Unified Inference Scaling via Online Joint Optimization of Model Routing and Test-Time Scaling
Kaiyu Huang, Xingyu Wang, Mingze Kong et al. (+6 more)
In real-world deployments of large
language models (LLMs), balancing inference quality and computational cost has become a central challenge. Existing approaches tackle this trade-off along two largely independent dimensions: model routing, which switches among models of different scales to match request complexity, and test-time scaling (TTS), which adjusts inference-time compute within a fixed model for fine-grained control. However, this decoupled design introduces inherent limitations. Model routing yields coarse-grained, discrete performance changes due to the sparse set of model scales, while single-model TTS often encounters capacity ceilings and exhibits diminishing returns as compute increases. Moreover, treating the two mechanisms separately restricts adaptability in dynamic inference environments. To overcome these limitations, we introduce Unified Inference Scaling (UIS), which unifies model routing and TTS in a single optimization space. Building on this formulation, we propose UniScale, an online framework that models adaptive UIS as a contextual multi-armed bandit problem and learns inference policies via LinUCB. The framework incorporates efficiency-aware learning and cost modeling to ensure stable and scalable optimization over high-dimensional action spaces. Evaluation shows that UniScale effectively exploits the synergy in the UIS space to deliver a fine-grained and consistently better quality-cost trade-off across diverse, dynamic inference scenarios.
Read the full abstract ▾
All authors Kaiyu Huang, Xingyu Wang, Mingze Kong, Zhubo Shi, Yuqian Hou, Hong Xu, Zhongxiang Dai, Minchen Yu, Qingjiang Shi
Computation and Language
The Flip Side of RLHF: On-Policy Feedback for Reward Model Self-Supervised Improvement
Xiaobo Wang, Tong Wu, Min Tang et al. (+3 more)
Building strong reward models (RMs) for
language model
alignment is bottlenecked by the cost and difficulty of acquiring diverse and reliable preference data from human annotation or judge models. It is dramatically worse as the policy evolves beyond the static RM training. Therefore, we propose SAVE (Self-supervised reward model improvement via Value-Anchored On-policy feedback), a framework that grades on-policy responses as feedback by using the value function for on-policy RM training. SAVE naturally converts the reward-graded on-policy responses into supervision with a
prompt-specific value head as an adaptive anchor. It computes RM advantages and filters ambiguous samples to update the RM via a contrastive objective. The effectiveness of SAVE for enhancing RM training is strongly validated through rigorous empirical evaluation across six diverse
benchmarks. It achieves outperforming results across all
datasets while maintaining consistent improvements across three RL
algorithms (GRPO, RLOO, GSPO) and different policy backbones.
Read the full abstract ▾
All authors Xiaobo Wang, Tong Wu, Min Tang, Jiaqi Li, Qi Liu, Zilong Zheng
Computation and Language
PatchWorld: Gradient-Free Optimization of Executable World Models for Agent Environments
Jiaxin Bai, Yue Guo, Yifei Dong et al. (+13 more)
World models for interactive text
agents must typically be learned from observation-action trajectories alone. Specifically, the environment returns text observations after each action, but does not expose a ground-truth latent state nor an inspectable transition model.A research gap remains in how to induce executable code as a world model in this black-box setting for prediction and
agent decision making. We introduce PatchWorld, a gradient-free framework that turns offline trajectories into executable Python world models through counterexample-guided code repair.Instead of predicting the next observation with a black-box model, PatchWorld induces symbolic belief-state programs whose action updates can be inspected, replayed, and locally patched. Across seven AgentGym environments, PatchWorld-Simple achieves the highest code-based decision-making score among evaluated methods (76.4% macro success in live one-step lookahead), matching or exceeding LLM-based lookahead while invoking no LLM calls inside the world-model prediction module itself. We further find that a human-specified residual-
memory bias improves surface observation fidelity but weakens
agent decision-making utility. This reveals a tradeoff in executable world models, since improving observation fidelity can come at the expense of action-discriminative dynamics, and vice versa. Code is available at https://github.com/HKBU-KnowComp/PatchWorld.
Read the full abstract ▾
All authors Jiaxin Bai, Yue Guo, Yifei Dong, Jiaxuan Xiong, Tianshi Zheng, Yixia Li, Tianqing Fang, Yufei Li, Yisen Gao, Haoyu Huang, Zhongwei Xie, Hong Ting Tsang, Zihao Wang, Lihui Liu, Jeff Z. Pan, Yangqiu Song
Computation and Language
dMoE: dLLMs with Learnable Block Experts
Sicheng Feng, Zigeng Chen, Gongfan Fang et al. (+2 more)
Diffusion Large
Language Models (dLLMs) have recently emerged as a promising alternative to autoregressive models, offering competitive performance while naturally supporting parallel decoding. However, as dLLMs are increasingly integrated with Mixture-of-Experts (MoE) architectures to scale model capacity, a fundamental mismatch arises between block parallel decoding and
token-level expert selection. Specifically, each dLLM forward pass processes multiple
tokens with bidirectional dependencies, whereas conventional MoE layers route each
token independently. This mismatch substantially increases the number of uniquely activated experts, making inference increasingly
memory-bound. To address this, we propose dMoE, a simple yet effective block-level MoE framework. The central idea of dMoE is to aggregate
token-level expert distributions within each block into a unified block-level expert distribution, which is then used to guide expert routing in a more coherent manner. In this way, dMoE substantially reduces the number of uniquely activated experts during inference without sacrificing performance, thereby mitigating the
memory-bound bottleneck. Extensive experiments across a variety of
benchmarks demonstrate the effectiveness of dMoE. On average, dMoE reduces the number of uniquely activated experts from 69.5 to 14.6 while retaining 99.11% of the original performance. Meanwhile, it reduces
memory usage by 76.64% to 79.84% and achieves 1.14$\times$ to 1.66$\times$ end-to-end latency speedup. Code is available at: https://github.com/fscdc/dMoE
Read the full abstract ▾
All authors Sicheng Feng, Zigeng Chen, Gongfan Fang, Xinyin Ma, Xinchao Wang
Computation and Language
MADS: Model-Aware Diverse Core Set Selection for Instruction Tuning
Yi Bai, Wenhao Zhang, Yao Chen et al. (+3 more)
Instruction fine-tuning is employed to enhance the instruction-following ability of large
language models (LLMs). As the amount of instruction fine-tuning data increases, selecting the optimal core set becomes particularly important. However, ensuring the diversity of the core set remains a significant challenge. Existing methods predominantly distinguish different training data based on the text features themselves, decoupled from LLMs' own understanding and representation of the data. To address this issue, we propose a Model-Aware Diverse Core Set Selection method, which distinguishes data features based on the neural activation states during LLM inference. This approach serves as an efficient instantiation of coverage-based selection using model-intrinsic activation features to ensure the diversity in the core set. We extensively evaluate our method on six
benchmarks that cover five distinct tasks. In our method, the core set selected by the 3B-parameter LLM performs effectively when utilized to fine-tune larger models with 7B, 8B, and 13B parameters. Experimental results on the Alpaca-GPT4
dataset, which comprises 52K instruction-response pairs, show that the core set, sized at 15% of the original
dataset and selected by Llama-3.2-3B-Instruct, achieves an average improvement of 2.5% when fine-tuning four larger base models compared with training on the full
dataset. The experimental results demonstrate that our method enhances model performance on multiple downstream tasks while reducing data requirements.
Read the full abstract ▾
All authors Yi Bai, Wenhao Zhang, Yao Chen, Jiao Xue, Zhumin Chen, Pengjie Ren
Computation and Language
Speculative Pipeline Decoding: Higher-Accuracy Drafting with Hidden Latency via Pipeline Parallelism
Yijiong Yu, Huazheng Wang, Shuai Yuan et al. (+2 more)
Speculative Decoding (SD) accelerates low-concurrency LLM inference with a draft-then-verify paradigm. Mainstream methods, however, rely on multi-
token prediction, which incurs compounding prediction difficulty and exposed draft latency. We propose Speculative Pipeline Decoding (SPD), which partitions the target LLM into $n$ pipeline stages so that $n$
tokens of a single sequence advance in parallel. To keep the pipeline saturated, a Pipeline Draft Module (PDM) aggregates multi-depth target features to predict the next
token and runs concurrently with each pipeline step, yielding bounded prediction difficulty, higher acceptance, and hidden draft latency. Experiments show that SPD achieves higher theoretical and wall-clock speedup than EAGLE-3 at moderate pipeline width, while more aggressive widths still leave room for further gains. Our code is available at https://github.com/yuyijiong/speculative_pipeline_decoding
Read the full abstract ▾
All authors Yijiong Yu, Huazheng Wang, Shuai Yuan, Ruilong Ren, Ji Pei
Computation and Language · Cross-listed from cs.CR
LLM Anonymization Against Agentic Re-Identification
Ziwen Li, Jianing Wen and Tianshi Li
Agentic LLMs with web search change the threat model for text anonymization: weak contextual cues can become cross-referenceable evidence for re-identification, yet those same details also carry downstream analytic value of the text. Existing defenses either remove explicit identifiers, perturb text for formal
privacy, or test rewritten text against non-web inference models, leaving underexplored the operating region between resistance to agentic web-search re-identification and utility retention. We introduce AURA (Anonymization with Utility-Retention Adaptation), an LLM-powered mask-reconstruct framework that decouples
privacy localization from utility-preserving reconstruction and selects candidates with
adversarial privacy and utility-retention checks. We evaluate AURA on real-user interview transcripts using re-identification attacks carried out by web-search
agents, along with a utility evaluation based on interviewee-profile facts, codebook facts, and the joint contextual utility grid. Our results show that adaptive-scope AURA yields the lowest agentic re-identification counts under each of three attacker models among the non-DP methods, and that at matched scope and backbone, AURA's mask-reconstruct design retains more contextual utility than the prior LLM anonymizer (+6.4 pp unit-grid recovery) at comparable
privacy. Source Code: https://github.com/AaronLi43/AURA
Read the full abstract ▾
Computation and Language
Fine-Tuning Improves Information Conveyance in Language Models
Yuwei Cheng, Weiyi Tian and Haifeng Xu
Fine-tuning is often believed to reduce uncertainty and diversity in large
language models, but existing analyses overlook output length, a key confounder, and therefore fail to capture how uncertainty is distributed across an entire generation rollout. To address this, we propose Canopy
Entropy ($\mathrm{CE}^\star$), a measure that views
language generation from a tree perspective, where “canopy” represents the space of all possible rollouts, making $\mathrm{CE}^\star$ naturally quantify the effective size of the generation space. $\mathrm{CE}^\star$ jointly captures uncertainty in both the output length $N$ and the generated sequence $Y_{1:N}$ – indeed, we show that it equals to total Shannon
entropy $H(N, Y_{1:N}\mid X)$, where $X$ denotes the
prompt. This formulation yields interpretable metrics, including a length-
entropy correlation term $\rho(N, r_N)$, where $r_N$ is the
entropy rate, quantifying information conveyance efficiency by indicating whether longer outputs are more or less informative per
token. Empirically, across tasks and model families, we find that fine-tuned models consistently exhibit stronger positive correlation $\rho(N, r_N)$, even when total
entropy decreases. Furthermore, after controlling for model family, task,
prompt, and output-length effects, we find that fine-tuning nearly triples the correlation strength between
entropy rate and semantic diversity, suggesting that aligned models convert
token uncertainty into semantic diversity more efficiently. Overall, these results demonstrate that fine-tuning does not simply reduce uncertainty, but fundamentally reorganizes it into more informative and semantically meaningful generations. Our code is available at https://github.com/WeiyiTian/canopy-
entropy.
Read the full abstract ▾
Computation and Language
Your Teacher Can't Help You Here: Combating Supervision Fidelity Decay in On-Policy Distillation
Yanjiang Liu, Jie Lou, Xinyan Guan et al. (+7 more)
On-policy
distillation transfers
reasoning capabilities by training a student model on its own generated trajectories using
token-level feedback from a teacher. However, we identify a critical bottleneck, Supervision Fidelity Decay (SFD): as student-generated prefixes lengthen, the teacher's next-
token distribution becomes less confident and less discriminative. Consequently, the teacher-dependent corrective signal in reverse-KL
distillation weakens, causing student drift to compound across long
reasoning chains. To mitigate SFD, we introduce \textbf{Lookahead Group Reward (\ours{})}. Building on the insight that next-step teacher confidence reflects the discriminative strength of future reverse-KL supervision, \ours{} evaluates the student's top-K candidate
tokens by the teacher confidence they induce at the subsequent step and assigns a group-normalized reward. To maintain computational efficiency, we further design an
entropy-triggered tree-
attention mechanism. Across six math and code
benchmarks, \ours{} improves mean@8 by 2.57 points over OPD for a 7B student, with gains increasing in longer-generation and reaching +4.92 points on AIME-26 at 39k
tokens.
Read the full abstract ▾
All authors Yanjiang Liu, Jie Lou, Xinyan Guan, Yuqiu Ji, Hongyu Lin, Ben He, Xianpei Han, Le Sun, Xing Yu, Yaojie Lu
Computation and Language
Beyond Agreement: Scoring Panel-Surfaced Biomedical Entity Candidates for Curator Triage
Shuheng Cao, Ruiqi Chen, Renjie Cao et al. (+3 more)
Biomedical NER is deceptively simple for modern LLMs: plausible biomedical mentions are easy to surface, but corpus-convention correctness depends on annotation conventions, span boundaries, entity granularity, and type schemas. Multi-LLM agreement is a salience signal, not corpus-convention correctness. We introduce a candidate-level panel-output
benchmark for panel-surfaced candidate verification, where the unit is an aligned candidate surfaced by an explicitly defined multi-model panel rather than a standalone extractor output. The
benchmark aligns eight LLMs' predictions over five public biomedical NER
datasets into a candidate master table. BioConCal is an in-domain supervised scorer that instantiates this layer with inference-time gold-free agreement, mention, surface-availability, and document features for a fixed candidate stream. In domain, BioConCal improves AUROC from 0.753 for raw agreement to 0.910. At a validation-selected 0.95 precision target it selects 1,340 candidates at empirical test precision 0.939, compared with 293 for raw agreement. This corresponds to candidate-level recall 0.592 and corpus-level recall 0.523 against a within-panel row-label ceiling of 0.883. The main benefit is not recovering entities missed by every panel member, but reshaping a noisy panel stream into a higher-yield review queue. Under entity-type shift, thresholds require target-domain validation, and exact character localization remains a separate deterministic post-processing step.
Read the full abstract ▾
All authors Shuheng Cao, Ruiqi Chen, Renjie Cao, Zhenhao Zhang, Siyu Zhang, Tingting Dan
Computation and Language
Incremental BPE Tokenization
Shenghu Jiang and Ruihao Gong
We propose a novel
algorithm for incremental Byte Pair Encoding (BPE) tokenization. The
algorithm processes each input byte in worst-case $\mathcal{O}(\log^2 t)$ time, leading to an overall complexity of $\mathcal{O}(n \log^2 t)$, where $n$ is the input length and $t$ is the maximum
token length. The
algorithm incrementally maintains BPE tokenization results for every prefix of the input text, implementing the standard BPE merge procedure defined by a fixed set of merge rules. This enables efficient partial tokenization in streaming settings. Functioning as a drop-in replacement for standard BPE, our approach achieves a speedup of up to ${\sim}3\times$ over Hugging Face's tokenizers, and demonstrates significant latency reductions over OpenAI's tiktoken on pathological inputs. We further introduce an eager output
algorithm that enables streaming output, emitting
tokens as soon as
token boundaries are determined during incremental tokenization. Overall, our results demonstrate that BPE tokenization can be performed incrementally with strong worst-case guarantees, while providing practical latency benefits in modern large
language model pipelines. Code: https://github.com/ModelTC/mtc-inc-bpe
Read the full abstract ▾
Computation and Language
Anchoring LLM Gender Bias to Human Baselines: A Cross-Lingual Audit
Jiwoo Choi, Seonwoo Ahn, Tongxin Zhang et al. (+1 more)
We audit six large
language models (LLMs) for gender stereotyping across English, Korean, Chinese, and Japanese. Three were developed primarily for English-
language use (Claude, GPT, Gemini) and three for East Asian use (DeepSeek, Syn-Pro, HyperCLOVA X). We adopt the HEXACO-100 personality inventory and anchor each model against a cross-cultural human
dataset spanning 48 countries to ask not whether LLMs are biased, but how far their gender attributions drift from the populations they are deployed among. Our findings show that their stereotyping spans a range roughly 2.5 times wider than the entire cross-country range found in humans, and the effect can compound across
languages. One English-centric model, prompted in Korean, reached 5 times the local baseline, even when the
prompt stated the candidate had already been hired, which often dampens human stereotyping. To characterize such behaviors without ranking them, we introduce a four-pattern framework – concordance, suppression, reorganization, and amplification – across 24 (model x
language)
cells. Item-level analysis reveals that translation does not just rescale stereotypes, but changes the attributes tied to it, hiding significant rearrangement under the surface while appearing well-calibrated. Our results ultimately suggest that no single debiasing pipeline is likely to address bias evenly across linguistic boundaries.
Read the full abstract ▾
All authors Jiwoo Choi, Seonwoo Ahn, Tongxin Zhang, Seohyon Jung
Computation and Language · Cross-listed from cs.IR
On the impact of retrieved content representations in RAG Pipelines
Jonathan J Ross, Bevan Koopman, Anton van der Vegt et al. (+1 more)
Retrieval-Augmented Generation (RAG) supplements a
language model's input with retrieved documents, yet most RAG pipelines inherit
retrieval components designed for human readers. How retrieved content should be represented when the consumer is a large
language model (LLM) rather than a human is less well understood. Recent work has proposed transformations of retrieved content and identified properties that affect generation, but each examines a single transformation or property in isolation, leaving open which features of a document's representation matter most. We address this with a controlled comparison: holding
retrieval fixed, we vary only the representation of retrieved documents, comparing an original baseline against thirteen transformations spanning selection, summarisation, and reformulation, in query-dependent and query-independent variants. Across these fourteen representations we measure question-answering accuracy for four generators, and for each representation we also measure answer retention: whether a known answer-bearing document still supports its answer after transformation. We find that answer retention is the primary determinant of generator accuracy; notably, when retention is high, a representation's wording, structure, length, and query-dependence have limited effect. This suggests that accuracy gains attributed to specific mechanisms in prior work may be partly explained by how well those mechanisms preserve answer-bearing content, an attribution that cannot be settled without controlling for retention.
Read the full abstract ▾
All authors Jonathan J Ross, Bevan Koopman, Anton van der Vegt, Guido Zuccon
Computation and Language
XLGoBench: Detecting cross-lingual skill gaps with algorithmic tasks
Purvam Jain, Preethi Jyothi, Vihari Piratla et al. (+1 more)
We introduce a set of synthetic algorithmic tasks to detect cross-lingual gaps in the abilities of large
language models. Our
benchmark is commensurate across
languages, since it requires models to perform the same underlying task in different
languages; scalable, since each task can be generated at varying levels of complexity allowing it to be adapted to models with different capabilities; quantifiable, since every task admits an objective notion of correctness; and transparent, since tasks are generated from simple templates that can be readily audited for translation errors. Because our
benchmark focuses on algorithmic tasks, differential performance is a sufficient – but not necessary – indicator of cross-lingual gaps. Nevertheless, we show through extensive experiments that our
benchmark exposes persistent cross-lingual gaps in multiple state-of-the-art models.
Read the full abstract ▾
All authors Purvam Jain, Preethi Jyothi, Vihari Piratla, Suvrat Raju
Computation and Language
Eywa: Provenance-Grounded Long-Term Memory for AI Agents
Resham Joshi
AI
agents that persist across sessions need
memory they can retrieve, audit, update, and erase. Existing
memory systems often collapse source evidence, extracted facts, retrieved context, and answer policy into one opaque
prompt path, making failures difficult to diagnose: a wrong answer may come from missing evidence, unsupported extraction, stale state,
retrieval loss, or answer-model behavior. We present Eywa, a provenance-grounded
memory architecture built around evidence before belief. Eywa stores immutable source evidence before deriving canonical facts, validates extracted
memories against typed signals and source support, and retrieves bounded
memory context through a deterministic multi-route read path with zero LLM calls inside
retrieval. Retrieved context is returned separately from answer instructions, allowing the same
memory substrate to be evaluated across frontier, budget, and local answer models. Under a frozen, artifact-recorded
retrieval configuration, Eywa reaches 90.19% judge accuracy on the LoCoMo C1-C4 split with Claude Sonnet 4.6 write and QA roles. On LongMemEval-S, it reaches 88.2%
retrieval-sufficiency accuracy. On BEAM, a 700-question technical-
memory stress
benchmark, it reaches 81.45% mean nugget score and 85.29% pass@score >= 0.5. Full per-question artifacts, including questions, gold answers, model answers, retrieved context, and labels, are published at https://eywa.to/research.
Read the full abstract ▾