Next point-of-interest (POI) recommendation aims to predict a user's next destination from spatio-temporal check-in data, and serves as the foundation for many location-based services. Recent advances incorporate large language models (LLMs) into this task by textualizing user trajectories to exploit their semantic reasoning ability. However, directly generating POI identifiers with such models is unreliable: POIs are sparse, localized, and long-tailed entities with weak natural-language semantics, which creates a semantic-to-entity gap between high-level mobility intent and concrete POI prediction. To bridge this gap, we propose ANCHOR, a category-anchored framework for semantic-to-entity grounding in next POI recommendation. Specifically, ANCHOR first estimates category-level mobility intent in a compact and stable semantic space, rather than predicting concrete POI entities directly. The inferred intent is then grounded into a candidate set by integrating three complementary sources of mobility evidence: category anchors, historical visit memory, and user preferences. Finally, a history-aware reranker selects the target POI by matching each candidate with the user's historical mobility patterns and current context. Experiments on three real-world check-in datasets show that ANCHOR consistently outperforms representative traditional, structure-enhanced, and LLM-based baselines on the reported metrics. Further analyses validate the effectiveness of semantic-to-entity grounding and show favorable trade-offs among candidate coverage, ranking quality, and inference efficiency.
We study satisficing regret minimization in bandits with knapsacks (BwK), where a learner sequentially selects randomized policies, i.e., distributions over arms, to obtain mean rewards above a satisficing threshold as frequently as possible while respecting finite resource budgets. The performance is measured by satisficing regret, the cumulative deficits of the selected policies' mean rewards below this threshold. Resource constraints introduce two difficulties: an algorithm can incur linear satisficing regret even with small standard regret as deficits to the threshold cannot be compensated, and random consumption can cause early resource depletion even when policies are feasible in expectation. To address these challenges, we propose SELECTwK, a general algorithmic template that uses a BwK learning oracle algorithm with sub-linear expected (standard) regret upper bound to search for and test candidate satisficing policies (i.e., policies whose mean rewards exceed satisficing threshold and whose expected resource consumption is feasible over the time horizon). Our regret analysis, supported by an LP-duality argument, establishes that SELECTwK attains constant expected satisficing regret when a satisficing policy exists. We further introduce a stronger regret condition that bounds the expected sum of positive reward deficits to the optimal one at each time step, under which SELECTwK achieves a sharper constant bound together with a sub-linear bound in the time horizon. As a complement, SELECTwK still preserves the oracle's sub-linear expected standard regret bound if the satisficing threshold is unattainable. Finally, we establish the oracle conditions and corresponding satisficing regret bound for existing finite-armed BwK and dynamic pricing algorithms, and conduct numerical experiments on dynamic pricing with limited inventory to demonstrate the empirical performance of SELECTwK.
Spatial transcriptomics (ST) enables joint characterization of tissue morphology and gene expression. However, 3D volumetric ST acquisition remains expensive. AI-driven methods offer a low-cost alternative but remain limited in accuracy, while recent 2D section-wise ST acquisition still rely on dense grid sampling, resulting in redundant measurements and undersampling of biologically informative regions. To bridge the gap, we propose STRATA, a reinforcement-driven active sampling framework with tri-paradigm adaptation for reconstructing dense 3D transcriptomic maps from sparse voxel-wise ST acquisition. STRATA constructs a 3D graph integrating H&E morphology with single-cell priors and employs a reinforcement learning agent to estimate morphological sparsity and transcriptomic informativeness, enabling budget-aware selection of informative spots. Given these sparse measurements, an adaptive tri-paradigm predictor combines retrieval, regression, and cross-layer imputation through lightweight fusion, dynamically weighting their contributions. Experiments across datasets demonstrate improved sampling efficiency and accuracy, particularly under low-budget settings. Code is available at https://github.com/Junchao-Zhu/STRATA.
Operational strategies for long-horizon robotic tasks require reliable evaluation to determine whether execution is progressing, stalling, regressing, or recovering from errors. Yet real- world robot datasets that capture error recovery and provide frame-level error annotations remain scarce. To address this gap, we introduce and open-source ReSTEP, a dataset of 3,156 episodes across two manipulation tasks with semantic stage labels and frame-level er- ror annotations. ReSTEP supports training and benchmarking progress monitors through successful execution, mistakes, and recovery. Existing progress estimators can struggle to distinguish visually similar task stages and maintain temporally consistent predictions while responding to genuine setbacks. We therefore propose REMORA, an online progress esti- mator that combines recurrent visual memory, distributional prediction, and causal peak selection. Memory retains execution history to help resolve visual ambiguity, while distribu- tional prediction represents alternative progress hypotheses and causal peak selection tracks a supported hypothesis over time. Together, these components enable consistent progress tracking that remains responsive to errors and recovery. Experiments on both ReSTEP tasks demonstrate that REMORA outperforms the state-of-the-art SARM baseline in progress ac- curacy and temporal consistency.
As egocentric video becomes increasingly available for embodied learning, an important research question is how to turn this low-cost, scalable source of data into robot manipulation capabilities. Egocentric videos naturally provide rich context for high-level task understanding, while real-robot demonstrations with wrist observations and action annotations provide supervision for fine-grained action alignment. We present Ego-CEVR, a data-native framework for hierarchical visual representation learning that assigns distinct learning objectives to global and interaction representations. The global module learns object relations, task state, and temporal progress through future prediction on egocentric video. The Interaction module interprets wrist observations in global context and learns local alignment and contact cues through robot action supervision. Separate residual paths expose both representations to the action policy, preserving direct access to task context alongside fine-grained execution information. Ego-CEVR supports independent pretraining of its global video module on large-scale human egocentric videos and videos from diverse robot embodiments, without requiring robot action labels or paired wrist observations.
Xin Huang, Teng Ma, Zhuoran Li, Yibo Zhao, Zhaoshi Li, Qianchao Zhu, Wei Yang, Chenxi Wang 等 10 位作者
KV cache reuse accelerates LLM serving, but existing systems manage host DRAM and NVMe SSD as a tiered hierarchy where DRAM acts as an intermediate cache. Production LLM workloads break both premises of this hierarchy. First, bimodal reuse defeats DRAM temporal filtering, since workloads are dominated either by one-shot blocks with long reuse distances or by multi-gigabyte agentic sessions that exceed DRAM capacity, putting SSD on the critical read path. Second, DRAM staging and eviction-ordered layout bottleneck SSD restoration. Holding host staging buffers for in-flight prefixes causes contention under load that aborts prefetches, while incremental eviction scatters a session's blocks across the SSD and prevents large sequential reads. We present FlatCake, a flat KV storage architecture that exposes host DRAM and NVMe SSD as peer media in a unified 64-bit address space. Exploiting the fact that KV blocks are write-once, immutable, and reused at session scope, FlatCake places each block directly onto its target medium in session-contiguous, layer-major order at creation time and never migrates it. FlatCake streams coalesced SSD reads directly into a bounded GPU window via GPUDirect Storage and decouples DRAM and SSD request queues to prevent head-of-line blocking. On a production Claude Code trace with DeepSeek-V4-Flash, FlatCake sustains a 94% cache hit rate as load increases (versus 64% for Mooncake), eliminates Mooncake's $6.5\times$--$6.7\times$ SSD read amplification, and reduces peak-load mean TTFT by $1.6\times$ while improving total token throughput by $1.6\times$. Across 8K--128K contexts on Llama-3.1-70B, FlatCake further reduces mean TTFT by up to $2.4\times$ over Mooncake and $1.8\times$ over Mooncake-GDS.
Existing code reveals past implementation decisions but rarely explains the reasoning behind them. When coding agents implement new features or modify existing functionality, past decisions may conflict with new task instructions. This raises a fundamental question for agentic software evolution: when multiple implementations satisfy the same functional requirements, how do coding agents decide whether to follow existing code or new task instructions? We investigate this question through bidirectional, controlled experiments in purpose-built repositories, followed by validation in real-world open-source repositories. Our approach isolates three competing sources of guidance: (1) agents' default preferences, (2) repository precedents, and (3) explicit task instructions. We conduct 5,625 trials across four coding agents and six implementation pairs, with validation in four open-source repositories spanning three programming languages. We find that agents share the same default implementation choices, yet overwhelmingly follow repository precedent when it conflicts with those defaults (359/360 trials). Crucially, not all precedents carry equal influence: three supporting local files yield 96% adoption, compared with only 4% for sibling-resource files, and changing file placement alone can shift adoption from 0% to 100% at a fixed example count. When new task intent conflicts with existing code, written priority rules achieve substantially higher success than code examples alone (59/60 versus 38/60 joint successes for two agents on two implementation pairs). The effectiveness of explicit rules also extends to open-source repositories, where a harness-loaded rule raises joint success from 0/120 to 117/120. These findings motivate design-decision traceability that preserves not only implementation choices but also the rationale and priorities that should guide future changes.
Modern image generators can produce complex visual artifacts, including diagrams, interfaces, and records that combine text, graphics, and layout to convey detailed information. They may refuse requests for overtly harmful imagery yet still generate artifacts that guide dangerous actions or support deception. We call the risk arising from what such images enable, rather than only what they depict, functional visual harm (FVH). Existing safety benchmarks focus largely on overtly harmful content, such as pornography and graphic violence, leaving image generators' safety alignment against FVH insufficiently evaluated. To address this gap, we introduce FVH-Bench, comprising 669 prompts across five harmful visual functions and ten risk domains. The 11 evaluated image generators rarely refuse these requests and frequently generate artifacts that fulfill the targeted harmful functions. High scores for target harm realization and harmful artifact usability further indicate that these outputs can support the intended harmful purposes. A further analysis finds that existing safety detectors also provide limited coverage of FVH, indicating gaps in their ability to recognize harmful visual functions. Together, these findings show that evaluating image-generator safety alignment requires considering not only what images depict, but also what they may enable.
Learning controllable world dynamics from heterogeneous interaction videos is challenging because robot embodiments use incompatible action spaces, while human videos lack native control labels. We present RynnWorld-Latent, a cross-embodiment video world model that uses the 3D-grounded latent actions introduced by RynnLAM as a shared conditioning interface. This representation enables joint pretraining on human interaction videos, robot demonstrations, and simulation rollouts without requiring a common native action space. Built on a pretrained video transformer, RynnWorld-Latent combines global action tokens with an identity-initialized local pathway that aligns actions with future video latents. We pretrain on 23 datasets comprising approximately 553.5M frames and adapt the model to target robots through post-training. The framework supports two data-generation routes: digital teleoperation uses target robot commands to generate action-video demonstrations, while latent-action synthesis transfers source motion to an edited initial frame and pairs the resulting rollout with retargeted pseudo robot actions. Experiments show improved video prediction metrics on public benchmarks and two real-robot platforms, alongside qualitative action following when different action sequences share the same initial observation. Synthetic mid-training improves policy success rates on LIBERO and LIBERO-Plus, and both generation routes improve real-world policy success rates over real-demonstration-only training in the evaluated settings. These results support universal latent actions as a scalable interface for learning controllable visual dynamics across heterogeneous embodiments.
Large language models (LLMs) can describe what a piece of code should compute, but they do so in free-form text. Such text cannot be checked against program executions, and it does not reveal whether a statement is correct. We present SemLoc, which turns LLM reasoning into executable, verifiable semantic constraints. With structured semantic grounding, the LLM writes each property as a constraint with a fixed category, attached to an anchor in the code graph, such as a variable definition, branch, loop, or return. SemLoc inserts each constraint as a runtime check and runs the tests, which yields a semantic violation spectrum over passing and failing tests. Counterfactual verification then patches the code at an anchor and reruns the tests; a patch that a constraint suggests counts as verified only if all tests then pass. We study SemLoc on fault localization, where the tests serve as the oracle for verification, with four LLMs on QuixBugs and on SemBugs, a new dataset of 809 real bugs from 750 GitHub projects. The constraints are executable: 99.4% become runtime checks, and on programs where coverage ties the faulty line with at least half of the covered lines, their violations alone rank it first in 40.9% of the cases, compared with 2.2% for spectrum-based fault localization and 7.0% for Daikon invariants. They make the LLM's answers verifiable: 68.4% of SemLoc's top-ranked answers carry a patch that the tests confirm, and 93.6% of these are correct, which the LLM's own confidence cannot match. SemLoc's first-rank accuracy does not differ significantly, for any of the four LLMs, from that of an LLM given the same inputs (81.4% vs. 82.0%), and it far exceeds that of an LLM that sees only the function (50.0%).
Web-based applications increasingly rely on understanding long multimodal documents, where answering questions often requires combining information across pages, tables, and figures. Agentic retrieval-augmented generation addresses these tasks through iterative retrieval and reasoning. However, Agentic execution is not always necessary, introducing additional computation and potentially incorrect answers. Our empirical analysis reveals that Agentic execution provides query-dependent benefits, while additional operations may improve, preserve, or degrade answer accuracy. These findings motivate controlling both Agent invocation and subsequent execution. To address these challenges, we propose RouteDRAG, a learning-based framework comprising four components. Unified Document Index organizes multimodal content into a graph-backed index and derives query representations from initial retrieval. Adaptive Path Routing preserves exact answering and predicts the relative benefit of Agent over shorter execution paths. Dynamic Agent Execution maintains an evolving workspace and refines Planner proposals to control tool use and termination. Sequential Policy Learning first trains operation control from observed outcomes and then trains path routing using the resulting Agent's performance. Across two benchmarks, RouteDRAG achieves an average accuracy improvement of 3.47 percentage points over shorter-path-only execution while assigning only 79.3% of queries to Agent. Notably, it also surpasses uniform Agent execution on both benchmarks while requiring fewer answering calls. Our code is now available at https://github.com/Ereshkigal1/RouteDRAG.
A skill is a modular package of natural-language instructions, executable scripts, and reference resources that an agent can load at runtime to extend its capabilities for a specific task. Skill-based agent systems therefore enable flexible reuse of third-party capabilities, but the openness of this skill ecosystem also opens up a new attack surface. Prior work has focused on vulnerabilities within individual skills, but little attention has been paid to risks that arise from interactions across skills. In this paper, we introduce skill cascading attacks, a threat paradigm in which a malicious objective is distributed across multiple skills so that each modification looks benign in isolation, yet their combined execution is harmful. For instance, in a prescription-review pipeline, the first skill weakens signals of recently discontinued medications in the extracted history, the second downgrades the severity of any drug interaction tied to them, and the third suppresses the resulting low-priority alert in the final summary, so that a severe drug-interaction warning silently disappears before reaching the physician. To systematically study this safety blind spot, we develop SkillCascade, an automated multi-agent red-teaming framework, and release SkillCascade-Bench, a benchmark of 213 validated cascading test cases across multiple agent systems and domains. Across representative agents (e.g., OpenClaw, Claude Code, Codex) and LLM backbones, cascaded interactions reliably induce harmful behaviors while evading existing per-skill scanners and runtime monitors. Our findings highlight a gap between component-level integrity and system-level safety, and call for defenses that reason over cross-skill interactions rather than individual skills in isolation.
Haonan Dong, Kehan Jiang, Haoran Ye Peking University, Wenhao Zhu, Zhaolu Kang, Guojie Song Peking University
Large Reasoning Models (LRMs) have recently achieved remarkable success in complex reasoning tasks. However, closer scrutiny reveals persistent failure modes compromising performance and cost: I) Intra-step level, marked by calculation or derivation errors; II) Inter-step level, involving oscillation and stagnation; and III) Instance level, causing maladaptive over-thinking. Existing endeavors target isolated levels without unification, while their black-box nature and reliance on RL hinder explainability and controllability. To bridge these gaps, we conduct an in-depth white-box analysis, identifying key neurons (Mixture of Neurons, MoN) and their fluctuation patterns associated with distinct failures. Building upon these insights, we propose NeuReasoner, an explainable, controllable, and unified reasoning framework driven by MoN. Technically, NeuReasoner integrates lightweight MLPs for failure detection with a special token-triggered self-correction mechanism learned via SFT. During inference, special tokens are inserted upon failure detection to actuate controllable remedial behaviors. Extensive evaluations across six benchmarks, six backbone models (8B 70B) against nine competitive baselines, demonstrate that NeuReasoner achieves performance gains of up to 27.0% while reducing token consumption by 19.6% 63.3%.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Minfeng Zhu Zhejiang University, Linxin Bao Zhejiang University, Wei Chen Zhejiang University, Linchao Zhu Zhejiang University
Visual Language Models (VLMs) have become a robust foundation for document question answering. Processing long documents remains challenging due to limited context windows and computational budgets. Existing page-level retrieval methods offer a practical solution, typically encoding pages and queries into vectors and ranking them via cosine similarity. However, such embedding-based methods (i) lack query–page interaction before similarity scoring and (ii) usually require large-scale datasets to align visual and textual embeddings. In this paper, we observe that the cross-modal attention maps of well-trained VLMs are able to highlight semantically relevant regions. Building on this insight, we present CAPS (Cross-modal Attention as Page Selector), a retrieval framework that utilizes attention mechanisms inside VLMs for page selection. Specifically, CAPS first enhances attention-based retrieval capability with a small amount of contrastive data, then identifies the most effective attention head through expert head selection, and finally employs an adaptive filtering mechanism to obtain an appropriate number of relevant page candidates. Extensive experiments on four long-document benchmarks demonstrate that CAPS outperforms state-of-the-art embedding-based methods in both retrieval precision and downstream DocQA accuracy. Notably, CAPS achieves these gains using less than 10% of the training data required by competing baselines, highlighting the data efficiency of attention-based page retrieval.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xue Jiang, Ge Li Peking University, Jiaru Qian, Xianjie Shi, Chenjie Li Peking University, Hao Zhu, Ziyu Wang Peking University, Jielun Zhang Peking University 等 14 位作者
Large language models (LLMs) excel at general programming but struggle with domain-specific software development. This gap motivates research into domain specialization methods that enable LLMs to learn and utilize domain knowledge and data. However, existing domain-specific code benchmarks focus on assessing what knowledge LLMs possess rather than how they acquire and apply new knowledge, lacking explicit knowledge corpora for developing domain specialization methods. To this end, we present KOCO-bench, a novel benchmark designed for evaluating domain specialization methods in real-world software development. KOCO-bench contains 6 emerging domains with 11 software frameworks and 25 projects, featuring curated knowledge corpora alongside multi-granularity evaluation tasks including domain code generation (from function-level to project-level with rigorous test suites) and domain knowledge understanding (via multiple-choice Q A). Unlike previous benchmarks that only provide test sets for direct evaluation, KOCO-bench requires acquiring and applying diverse domain knowledge (APIs, rules, constraints, etc.) from the corpora to solve evaluation tasks. Our evaluations reveal that KOCO-bench poses significant challenges to state-of-the-art LLMs. Even with domain specialization methods (e.g., SFT, RAG, kNN-LM) applied, improvements remain marginal. Best-performing coding agent, Claude Code, achieves only 34.2%, highlighting the urgent need for more effective domain specialization methods. We release KOCO-bench, evaluation code, and baselines to advance further research at https://github.com/jiangxxxue/KOCO-bench.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Yerong Wu Southeast University, Tianxing Wu, Minghao Zhu Alibaba Group, Hangyu Sha Southeast University, Haofen Wang Tongji University
Achieving realistic human-like conversation for virtual characters requires not only a simple memorization and recall of past events, but also the strategic utilization of memory to meet factual needs and social engagement. Current memory utilization relevant (e.g., memory-augmented generation, long-term dialogue, and etc.) benchmarks overlook this nuance, treating memory primarily as a static repository of facts rather than a dynamic resource to be strategically deployed in dialogues. To address this gap, we design StratMem-Bench, a new benchmark to evaluate strategic memory use in character-centric dialogues. This dataset comprises 657 instances where virtual characters must navigate heterogeneous memory pools containing required, supportive, and irrelevant memories. We also propose a framework with different evaluation metrics including Strict Memory Compliance, Memory Integration Quality, Proactive Enrichment Score and Conditional Irrelevance Rate, and to evaluate strategic memory use capabilities of virtual characters. Experiments on StratMem-Bench which leverage the state-of-the-art large language models as virtual characters show that all models perform well at distinguishing between required and irrelevant memories, but struggle once supportive memories are introduced into the decision process.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Hongxin Ding, Baixiang Huang, Yue Fang Peking University, Weibin Liao, Xinke Jiang, Jinyang Zhang, Yinghao Zhu The University of Hong Kong, Zheng Li 等 11 位作者
Interactive medical questioning is essential in clinical consultations, where physicians must actively gather necessary patient information. Yet existing medical Large Language Models (LLMs) predominantly follow a reactive paradigm, risking diagnostic errors by answering before seeking sufficient details. To bridge this gap, we propose ProMed, a reinforcement learning framework that transitions LLMs toward a proactive paradigm, enabling them to ask clinically valuable questions before decision-making. Central to ProMed is the Shapley Information Gain (SIG) reward, which quantifies a question’s clinical utility as the amount of newly acquired information, while considering its contextual importance via Shapley values. We integrate SIG into a two-stage training pipeline: (1) SIG-Guided Model Initialization uses Monte Carlo Tree Search to construct high-reward interaction trajectories for supervision, and (2) SIG-Augmented Policy Optimization, with a novel SIG-guided Reward Distribution Mechanism that prioritizes informative questions for fine-grained optimization. Experiments on partial-information medical benchmarks show that ProMed significantly outperforms state-of-the-art methods by 6.29% on average and delivers a 54.45% gain over the reactive paradigm, and generalizes robustly to out-of-domain cases. Our codes are available at https://github.com/hxxding/ProMed.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Cui Yakun The Hong Kong University of Science and Technology, Peng Qi Centre for Trusted Internet and Community, National University of Singapore, Fushuo Huo The Hong Kong University of Science and Technology, Hang Du Beijing University of Posts and Telecommunications, Weijie Shi Hong Kong University of Science and Technology, Juntao Dai Peking University, Zhenghao Zhu, Sirui Han The Hong Kong University of Science and Technology
The advent of multi-modal large language models (MLLMs) has greatly advanced research on video fake news detection (VFND) tasks. Existing benchmarks typically focus on the detection accuracy, while failing to provide fine-grained assessments for the entire detection process. To address these limitations, we introduce POVFNDB (Process-oriented Video Fake News Detection Benchmark), a process-oriented benchmark comprising 10 tasks designed to systematically evaluate MLLMs’ perception, understanding, and reasoning capabilities in VFND. This benchmark contains 36,240 human-annotated question-answer (QA) in structured or open-ended formats, spanning 15 distinct evaluation dimensions that characterize different aspects of the video fake news detection process.Using POVFNDB, we conduct comprehensive evaluations on both proprietary and open-source MLLMs. Moreover, We fine-tune Qwen2.5VL-7B-Instruct on a reasoning dataset generated by our proposed POVFND-CoT, a chain-of-thought method that utilizes rationales from evaluation results and rationale validation. The resulting model achieves sota performance on VFND.
尚无 AI 解读，展示原文摘要
cs.CL · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Siran Liu, Yang Ye, Qianchao Zhu, Zane Cao, Yongchao He
Autoregressive decoding inherently limits the inference throughput of Large Language Model (LLM) due to its sequential dependency. Speculative decoding mitigates this by verifying multiple predicted tokens in parallel, but its efficiency remains constrained by what we identify as verification heterogeneity—the uneven difficulty of verifying different speculative candidates. In practice, a small subset of high-confidence predictions accounts for most successful verifications, yet existing methods treat all candidates uniformly, leading to redundant computation. We present HeteroSpec, a heterogeneity-adaptive speculative decoding framework that allocates verification effort in proportion to candidate uncertainty. HeteroSpec estimates verification complexity using a lightweight entropy-based quantifier, partitions candidates via a data-driven stratification policy, and dynamically tunes speculative depth and pruning thresholds through coordinated optimization. Across five benchmarks and four LLMs, HeteroSpec delivers an average 4.24× decoding speedup over state-of-the-art methods such as EAGLE-3, while preserving exact output distributions. Crucially, HeteroSpec requires no model retraining and remains compatible with other inference optimizations, making it a practical direction for improving speculative decoding efficiency.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xinhe Shi, Qingcheng Zeng, Weihao Xuan The University of Tokyo and RIKEN, Linchao Zhu Zhejiang University
Large language models (LLMs) exhibit uneven performance across languages. In language-specific applications, practitioners often rely on target-language corpora or cross-lingual transfer to achieve better performance. However, traditional linguistic typology, commonly used as a transfer language selection strategy in previous studies, may not align with LLM’s perception of language similarity. This work proposes LLM-based language similarity as a novel perspective for selecting effective fine-tuning languages. We construct a framework to quantify the similarity within each language pair through both the lenses of language-specific performance patterns and cross-lingual transferability, ultimately deriving three similarity score matrices. Moreover, we observe a counter-intuitive phenomenon: super-additive transfer effect, where fine-tuning on a certain language yields higher performance than fine-tuning directly on the target language. Additionally, due to the absence of an existing dataset meeting our experimental requirements, we construct and release M4CQ-Pro dataset, which features domain-diverse distribution of 135 tasks and content consistency across 31 languages (including over 20 medium- and low-resource languages), with 61518 manually reviewed high-quality questions per language. We evaluate our approach on representative multilingual LLMs and results show that all three LLM-based similarity measures effectively guide fine-tuning language selection, outperforming traditional linguistic similarity, with the integrated measure achieving the best results. Our approach provides not only a novel perspective on language similarity, but also practical baselines for selecting fine-tuning languages.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xichen Zhang Hong Kong University of Science and Technology, Ziyi He, Yinghao Zhu The University of Hong Kong, Sitong Wu The Chinese University of Hong Kong, Shaozuo Yu, Meng Chu, Wenhu Zhang, Haoru Tan 等 9 位作者
Search agents have emerged as a pivotal paradigm for solving open-ended, knowledge-intensive reasoning tasks. However, training these agents via Reinforcement Learning (RL) faces a critical dilemma: interacting with live commercial Web APIs is prohibitively expensive, while relying on static data snapshots often introduces noise due to data misalignment. This misalignment generates corrupted reward signals that destabilize training by penalizing correct reasoning or rewarding hallucination. To address this, we propose SearchGym, a simulation environment designed to bootstrap robust search agents. SearchGym employs a rigorous generative pipeline to construct a verifiable knowledge graph and an aligned document corpus, ensuring that every reasoning task is factually grounded and strictly solvable. Building on this controllable environment, we introduce SearchGym-RL, a curriculum learning methodology that progressively optimizes agent policies through purified feedback, evolving from basic interactions to complex, long-horizon planning. Extensive experiments across the Llama and Qwen families demonstrate strong Sim-to-Real generalization. Notably, our Qwen2.5-7B-Base model trained within SearchGym surpasses the web-enhanced ASearcher baseline across nine diverse benchmarks by an average relative margin of 10.6%. Our results validate that high-fidelity simulation serves as a scalable and highly cost-effective methodology for developing capable search agents.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Lei Liu Alibaba Group, Hao Zhu Ant Group, Xiaoyan Yang, Yue Shen antgroup, Zhixuan Chu Zhejiang University, Jian Wang, Jinjie Gu, Kui Ren Shanghai Institute for Advanced Study, Zhejiang University, The State Key Laboratory of Blockchain and Data Security, Zhejiang University and Zhejiang University
Continual Pre-training (CPT) serves as a fundamental approach for adapting foundation models to domain-specific applications. Scaling laws for pre-training define a power-law relationship between dataset size and the test loss of an LLM. However, the marginal gains from simply increasing data for CPT diminish rapidly, yielding suboptimal data utilization and inefficient training. To address this challenge, we propose a novel perplexity-aware data scaling law to establish a predictive relationship between the perplexity landscape of domain-specific data and the test loss. Our approach leverages the pre-trained model’s own perplexity on domain data as a proxy for estimating the knowledge gap, effectively quantifying the informational perplexity landscape of candidate training samples. By fitting this scaling law across diverse perplexity regimes, we enable adaptive selection of high-utility data subsets, prioritizing content that maximizes knowledge absorption while minimizing redundancy and noise. Extensive experiments on both medical and general-domain benchmarks demonstrate that our method consistently identifies near-optimal training subsets, achieving superior performance with significantly reduced data consumption.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Shuyan Zhou, Frank F. Xu, Hao Zhu, Xuhui Zhou, Robert Lo, Abishek Sridhar, Xianyi Cheng, Tianyue Ou 等 12 位作者
With advances in generative AI, there is now potential for autonomous agents to manage daily tasks via natural language commands. However, current agents are primarily created and tested in simplified synthetic environments, leading to a disconnect with real-world scenarios. In this paper, we build an environment for language-guided agents that is highly realistic and reproducible. Specifically, we focus on agents that perform tasks on the web, and create an environment with fully functional websites from four common domains: e-commerce, social forum discussions, collaborative software development, and content management. Our environment is enriched with tools (e.g., a map) and external knowledge bases (e.g., user manuals) to encourage human-like task-solving. Building upon our environment, we release a set of benchmark tasks focusing on evaluating the functional correctness of task completions. The tasks in our benchmark are diverse, long-horizon, and designed to emulate tasks that humans routinely perform on the internet. We experiment with several baseline agents, integrating recent techniques such as reasoning before acting. The results demonstrate that solving complex tasks is challenging: our best GPT-4-based agent only achieves an end-to-end task success rate of 14.41%, significantly lower than the human performance of 78.24%. These results highlight the need for further development of robust agents, that current state-of-the-art large language models are far from perfect performance in these real-life tasks, and that WebArena can be used to measure such progress.
Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li 等 13 位作者
Evaluating large language model (LLM) based chat assistants is challenging due to their broad capabilities and the inadequacy of existing benchmarks in measuring human preferences. To address this, we explore using strong LLMs as judges to evaluate these models on more open-ended questions. We examine the usage and limitations of LLM-as-a-judge, including position, verbosity, and self-enhancement biases, as well as limited reasoning ability, and propose solutions to mitigate some of them. We then verify the agreement between LLM judges and human preferences by introducing two benchmarks: MT-bench, a multi-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our results reveal that strong LLM judges like GPT-4 can match both controlled and crowdsourced human preferences well, achieving over 80% agreement, the same level of agreement between humans. Hence, LLM-as-a-judge is a scalable and explainable way to approximate human preferences, which are otherwise very expensive to obtain. Additionally, we show our benchmark and traditional benchmarks complement each other by evaluating several variants of LLaMA and Vicuna. The MT-bench questions, 3K expert votes, and 30K conversations with human preferences are publicly available at https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.