arXiv
cs.CL
- 01TabletCraft: Bridging a 4,000-Year Cultural Gap with Bidirectional Akkadian NMT and Cuneiform RenderingarXiv:2608.02609v1 Announce Type: new Abstract: Half a million cuneiform clay tablets survive in museums worldwide, yet modern users can neither read nor write in the world's oldest writing system, leaving a 4,000-year cultural barrier that existing NLP tools have only partially addressed. Prior work enables one-way, scholar-oriented translation from Akkadian to English, but offers no path in the reverse direction: non-specialist users cannot compose new content in cuneiform, and therefore remaiZhaohui Wang
- 02BBOWP-Bench: Evaluating LLMs on Black-Box Optimization Word ProblemsarXiv:2608.02612v1 Announce Type: new Abstract: Formulating an optimization problem strongly affects the quality of the final solution, yet good formulations usually require substantial expertise. Recent studies have therefore examined how to automatically derive optimization problems from natural-language descriptions, but existing benchmarks focus on settings where objectives and constraints can be written explicitly as mathematical expressions. Many practically important problems are naturallYutaro Yamada, Kei Hiroshima, Nozomu Yoshinari, Kento Uchida, Shinichi Shirakawa
- 03MemArena: An Ego-Centric Benchmark for On-Device Agentic Personal Memory Assistants at ScalearXiv:2608.02613v1 Announce Type: new Abstract: Edge-deployed personal memory assistants must handle private interpersonal conversations on-device with open-weight models. Yet, existing memory benchmarks often under-test the combination of activity-dense interaction, ego-centric perspective, and coherent multi-session worlds. MemArena fills these gaps with a single-world conversational benchmark built with its MASim agent simulator, for 50 agents over 15 days (10.3M dialog-text tokens, 24.1K texJiadong Zhang, Xiaosong Ma
- 04OncoTriad-QA: A Patient-Level Radiology-Pathology-Genomics Benchmark for Pan-Cancer ReasoningarXiv:2608.02615v1 Announce Type: new Abstract: Cancer diagnosis and characterization require integrating complementary evidence from radiology, pathology, genomics, and clinical metadata. However, most medical large language model (LLM) and vision-language model (VLM) benchmarks focus on isolated modalities or narrow image-text tasks, leaving patient-level oncology assessment across multiple evidence streams largely untested. We introduce OncoTriad-QA, a patient-level radiology-pathology-genomiAhnaf Munir, Dannong Wang, Michael W. McDonald, Mubarak Shah, Pegah Khosravi, Yu Tian
- 05Evaluating OpenAI's Privacy Filter: Cross-Lingual, Cross-Domain PII Detection Across 42 BenchmarksarXiv:2608.02616v1 Announce Type: new Abstract: We present the first independent, systematic evaluation of OpenAI's Privacy Filter (OPF), a 1.5B-parameter bidirectional PII detector, across 42 synthetic benchmarks spanning 22 languages and 5 domains. Zero-shot, OPF achieves F1=0.855 on AI4Privacy and 0.464 on SPY medical, outperforming Presidio (0.431, 0.273) and XLM-RoBERTa (0.269, 0.111) on PII-annotated benchmarks; on multilingual NER, XLM-RoBERTa leads OPF on all 13 Indic and non-Latin languRohith Uppala
- 06Preferred, Not Safer: Pairwise Preference Is a Poor Proxy for Clinical SafetyarXiv:2608.02617v1 Announce Type: new Abstract: We evaluate whether clinician pairwise preferences provide a reliable signal of clinical safety in large language model (LLM) evaluation using expert feedback from MOOVE (Massive Open Online Validation and Evaluation), a clinician-led platform collecting blinded pairwise preferences alongside multi-criterion rubric ratings. Clinicians assign scores on a discrete $[-2, +2]$ scale, where negative values indicate clinically unsafe or misleading contenFay Elhassan, David Sasu, Alexandra Kulinkina, Lars Henning Klein, Mary-Anne Hartley
- 07JudgeArena: A Unified Framework for Reproducible LLM-Judge EvaluationarXiv:2608.02620v1 Announce Type: new Abstract: LLM-as-a-judge evaluation has become a dominant paradigm for ranking language models, yet the ecosystem remains fragmented: most benchmarks ship their own code base, hardcode a specific closed-model judge, and support a single evaluation protocol. This fragmentation makes it difficult to study how design choices--the benchmark, the judge model, the prompt, the inference backend--affect the conclusions we draw about model quality. We introduce JudgeErlis Lushtaku, Bora Kargi, Ali Elganzory, Fabio Ferreira, Alejandro R. Salamanca, Julia Kreutzer, David Salinas
- 08Knowing the Form, Not the Function: Automatically Auditing Answer--Authority Decoupling in Legal BenchmarksarXiv:2608.02621v1 Announce Type: new Abstract: Legal benchmarks typically score final answers even when models also state legal authority. We test whether answer correctness can serve as a proxy for authority grounding. Under ordinary reasoning prompts that did not request statutory citations, four LLMs spontaneously produced authority markers across 238 Taiwan bar-examination items. Because each item has a verified governing provision, we automatically audit answer correctness and authority grHsien-Jyh Liao
- 09Speculative Correction: Draft-then-Refine Decoding for Diffusion Language ModelsarXiv:2608.02625v1 Announce Type: new Abstract: Diffusion language models (DLMs) can revise tokens bidirectionally, but standard decoding procedures often adapt them to left-to-right generation by producing text block by block. We study a simple plug-and-play inference pattern: first generate a complete draft, then refine the full response using bidirectional diffusion. Using LLaDA2.1-Flash and LLaDA2.1-Mini, we evaluate two configurations. In Flash-Flash, the same Flash model serves as both draBrian K Chen, Chong Wu, Kenji Kawaguchi
- 10Stuck on "A": Diagnosing and Repairing Interface Injury in Attention-to-KDA Linearization of a 0.6B Language ModelarXiv:2608.02689v1 Announce Type: new Abstract: We convert 21 of 28 full-attention layers of Qwen3-0.6B-Base into KDA (Kimi Delta Attention) linear-attention layers on a single consumer-grade GPU budget, and ask a simple question: what exactly does the conversion break? After surgery, hidden-state alignment and end-to-end KL distillation drive the student close to its teacher in perplexity, yet multiple-choice accuracy stays near random chance (25-29% vs. the teacher's 50.6% on C-Eval). Using aRonglong Bao
- 11Crayotter: Learning Long-Horizon Video Editing Agents via Group-Relative Preference BackpropagationarXiv:2608.02694v1 Announce Type: new Abstract: Long-horizon video editing agents receive final-product feedback only after many interdependent decisions. Yet editing quality is subjective, admits multiple valid solutions, and is not meaningfully calibrated across heterogeneous requests, making a global scalar objective both ambiguous and temporally uninformative. Our key observation is that fixing the request, materials, and production constraints converts this subjective objective into an ordiLecheng Yan, Jianze Lin, Yichong Zhang, Ben Pan, Wenxi Li, Chenyang Lyu, Liting Zhou, Cathal Gurrin
- 12ARCHead: Activation-Metric Residual Correction for Large Language Model Output HeadsarXiv:2608.02703v1 Announce Type: new Abstract: Weight-only quantization substantially reduces the storage of large language model (LLM) transformer blocks, but practical backends often retain the final language-modeling head (LM-head) in BF16 or FP16. Quantizing this projection naively can strongly perturb the vocabulary-logit distribution. We present ARCHead, a packed LM-head compressor that combines a quantized low-rank core, group-wise INT4 residuals, and a low-rank correction fitted in an a\c{S}uayp Talha Kocabay, Talha R\"uzgar Akku\c{s}, Kamer Ali Yuksel
- 13Learning a Vector-Symbolic Model for Socio-Cultural TasksarXiv:2608.02807v1 Announce Type: new Abstract: How can we better represent the impact of sociocultural structures on decision making in computational cognitive models? Modeling this impact requires traversing multiple levels of semantic representation, however it is not immediately clear to a modeler which levels of representation are most salient to a given situation. Though large language models and cognitively grounded corpus models can represent broad semantic associations through co-occureMeera Ray, Swapnika Dulam, Christopher L. Dancy
- 14BODHI: Do LLMs Branch Out and Discover Heterogeneous Inferences?arXiv:2608.02867v1 Announce Type: new Abstract: Although reinforcement learning with verifiable rewards (RLVR) has improved the performance of large language models (LLMs) across a variety of reasoning tasks, there is significant debate as to whether RLVR expands the reasoning capability boundary, or just improves sampling efficiency. In this paper, we investigate the nature of test-time exploration in RLVR-trained LLMs by employing controlled maze-solving experiments and extracting a tree strucSoumadeep Saha, Krish Sharma, Akshay Chaturvedi, Nicholas Asher
- 15FLARE: Few-shot Learning-based Adaptive Reflective EnginearXiv:2608.02919v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in complex, compound AI systems where performance hinges on the quality of prompts. Recent state-of-the-art optimizers like GEPA (Genetic-Pareto) have argued that reflective instruction evolution can outperform traditional reinforcement learning and few-shot optimization. In this work, we challenge this shift by introducing FLARE (Few-shot Learning-based Adaptive Reflective Engine), a frameworkDhanasekar Sundararaman, Bharat Gandhi, Aashna Garg, Minjie Li
- 16Character Iconicity vs. Arbitrariness: An Arabic NLP PerspectivearXiv:2608.02935v1 Announce Type: new Abstract: Arabic script uses 28 letters, many of which share a common base shape (rasm) and are distinguished only by dot placement. Because early Arabic manuscripts were written without dots yet remained interpretable, dot removal offers a natural test of whether these visual distinctions are functionally necessary. Prior work has shown that dotless Arabic can remain readable and effective for natural language processing (NLP), but it remains unclear whetheDorieh Alomari, Irfan Ahmad, Maged S. Al-shaibani
- 17Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory SpeecharXiv:2608.02941v1 Announce Type: new Abstract: We audit five frontier large language models on native Bangla derogatory speech (gali) across six protocols to test a single hypothesis: Comprehension-Containment Decoupling. We propose that contemporary safety alignment is bound to high-resource surface forms rather than harmful meaning, causing a model's capacity to comprehend a low-resource slur and its capacity to contain it to operate independently. Every protocol corroborates this hypothesisShadab Bin Habib, A K M Ferdous Reza Habib, Subarno Neel, Adib Sakhawat
- 18OPTD: On-Policy Transition Distillation with Consistency-Guided Adaptive Compression for Few-Step Diffusion Language ModelsarXiv:2608.02942v1 Announce Type: new Abstract: Diffusion language models (dLLMs) can predict many tokens in parallel, but accurate generation still requires many iterative denoising steps. Few-step distillation accelerates decoding by compressing multiple teacher steps into a single student transition. However, existing methods construct supervision on off-policy trajectories. At inference, the student's early parallel commitments alter the context of later predictions, so the states it actuallXiaocheng Lu, Hualei Zhang, Shuhan Guo, Jie Zhang, Xiaoyi Pang, Jian Liu, Haoxi Li, Bohai Gu, Haoxuan Che, Jingcai Guo, Song Guo
- 19Every Wrong Answer Counts: Option-Level Psychometrics for LLM Multiple-Choice BenchmarksarXiv:2608.02966v1 Announce Type: new Abstract: Most multiple-choice question (MCQ) benchmarks evaluate Large Language Models (LLMs) only by whether they select the correct answers. This binary scoring treats all incorrect responses alike, even though an LLM's preferences among incorrect options may contain systematic and useful information about its behavior and ability. We introduce the LLM Nominal Response Model (LLM-NRM), an option-aware psychometric framework that models the full distributiXiao Fei, Yang Zhang, Sarah Almeida Carneiro, Michalis Vazirgiannis
- 20Mapping the City Through the Lens of Language ModelsarXiv:2608.02971v1 Announce Type: new Abstract: Language models often complete an underspecified reference to a city with unstated assumptions about urban size, form, infrastructure, environment, and function. We measure those assumptions without naming places. Ten open-weight checkpoints rate anonymized profiles derived from real morphological urban centres across 40 audited indicators and seven domains. The design combines constrained probability-based ratings, prespecified reliability screensWanqi Liu, Rong Zhao, Zhizhou Sha, Qinyu Cui, Yecheng Zhang
- 21TQLite: Multi-LLM Jury Guided Distillation for Real-time MQM Translation Quality EvaluationarXiv:2608.02975v1 Announce Type: new Abstract: Large language models (LLMs) have demonstrated impressive performance in MQM-based translation quality (TQ) evaluation, and recent advances in large reasoning models (LRMs) promise even greater improvements. However, both LLMs and LRMs are computationally expensive to deploy at scale, while small language models (SLMs)---though much more efficient---struggle with the complex reasoning required for evaluation tasks. In this work, we present an extenBhavin Jawade, Cameron R. Wolfe
- 22On the Non-Specificity of Statistical Measures Used in Script DeciphermentarXiv:2608.02999v1 Announce Type: new Abstract: Statistical regularities are routinely offered as evidence that undeciphered sign systems encode language; the Indus script debate is the canonical example. Any such inference rests on specificity: the reported outcome must be unusual among plausible structured non-languages. We test that premise constructively with SIGIL, a purpose-built generative emblem system whose 3,000-text core corpus carries explicit compositional meanings although no signNikhil Raghavendra
- 23Language Models Encode the Contextual Truth of PropositionsarXiv:2608.03035v1 Announce Type: new Abstract: Prior work has shown that LLMs encode the truth of factual propositions along linear directions in activation space. It's unclear how these representations extend to contextual truth: propositions whose truth is determined by in-context evidence rather than world knowledge. We show that LLMs maintain a linear representation of contextual truth that persists across structurally different output policies, even when the output doesn't require the modeRupak Sarkar, Pritika Ramu, Rachel Rudinger
- 24Beyond Accuracy: A Multidimensional Evaluation of Statistical Reasoning in Large Language ModelsarXiv:2608.03038v1 Announce Type: new Abstract: Statistical reasoning is multidimensional, yet evaluations of large language models (LLMs) typically emphasize response accuracy while overlooking how models construct and communicate statistical explanations. This study demonstrates the value of a multidimensional evaluation by combining response accuracy, response behavior, structural topic modeling, and lexical similarity analysis. The framework is applied to explanations generated by 15 currentMonnie McGee, Mateo Langston Smith, Julian Cabrera
- 25Emulate or Estimate? The Divergent Strengths of Base and Post-Trained Language Models for Opinion SimulationarXiv:2608.03044v1 Announce Type: new Abstract: Large language models are increasingly used to simulate human opinions, but prior work reports conflicting results: some studies find promising alignment with human survey data, while others find persona collapse and weak demographic sensitivity. We show that much of this conflict stems from conflating two distinct tasks. We call the first task emulation, in which models generate individual responses that aggregate into a population distribution. WSeth Grief-Albert, Jessica Bo, Difan Jiao, Ashton Anderson
- 26PI-Mem: Pushing Long-Context Reasoning to 3.6M Tokens with Parallel-Iterative MemoryarXiv:2608.03048v1 Announce Type: new Abstract: Long-context reasoning remains a critical bottleneck for large language models, as recent recurrent-memory approaches face two inherent challenges: sequential chunk-wise updates can overwrite early critical evidence with later irrelevant content, and serial inter-chunk dependencies limit parallelism and cause latency to increase with context length. To address these issues, we propose PI-Mem (Parallel-Iterative Memory), a mechanism that processes aDawei Liu, Haixu Song, Shuang Cheng, Shijie Wang, Haozheng Hou, Kaifeng Liu, Ermo Hua, Zhonghang Yuan, Zhijie Zhong, Yuchen Fan, Biqing Qi, Bowen Zhou
- 27SeqLLM: Augmenting LLMs with Behavioral-Sequence Modeling for High-Stakes Decisions at WeChat PayarXiv:2608.03063v1 Announce Type: new Abstract: Merchant risk control at large payment platforms screens tens of millions of merchants daily, where false positives harm legitimate merchants and false negatives leave harmful activity undetected. The hardest cases require jointly understanding a merchant's textual profile and long behavioral sequence. Large language models (LLMs) excel at text but cannot natively model such sequences, while adapting them often causes catastrophic forgetting. We prGuilin Li, Jiaxing Zhang, Matthias Hwai Yong Tan, Bo Wang, Weiran Huang
- 28Activation-Guided Neuron Intervention to Induce Alzheimer's-Related Computational Language Phenotypes in a Large Language ModelarXiv:2608.03067v1 Announce Type: new Abstract: Changes in spontaneous speech provide an early signal of cognitive dysfunction in Alzheimer's disease (AD) that large language models (LLMs) can detect. However, detection alone cannot establish whether the underlying model representations contribute functionally to behavior. We introduce an activation-guided intervention framework using Qwen3-8B. The framework identifies feed-forward neurons with higher activation rates for AD than control transcrRui He, Ercong Nie, Hong Jiang, Iris E. Sommer, Philipp Homan, Wolfram Hinzen
- 29CVPO: Enhancing LLM Reinforcement Learning Reasoning via Value-Variance Adaptation and Dynamic Curriculum LearningarXiv:2608.03068v1 Announce Type: new Abstract: Reinforcement learning (RL) has emerged as an effective method for enhancing the reasoning capabilities of large language models (LLMs). However, existing methods suffer from insufficient precision in feedback on generated answer trajectories and exhibit the phenomenon of problem difficulty drift. To address these challenges, we propose CVPO - Curriculum-guided Value-Variance Policy Optimization. At the response trajectory level, we find that tokenZiqi Jia, Yalu Ouyang, Bo Pang, Panpan Li, Hangfei Xu, Shengzhao Wen, Shiyong Li, Yanpeng Wang
- 30PAMT: Process-Aligned Reinforcement Learning for Multi-Domain Machine TranslationarXiv:2608.03077v1 Announce Type: new Abstract: Multi-domain machine translation (MDMT) requires more than fluent generation: it demands domain-sensitive translation decisions such as domain disambiguation, terminology control, and stylistic adaptation. Large reasoning models (LRMs) make such decisions explicit through intermediate translation steps, but our analysis across 15 domains and four translation directions shows that this explicit reasoning is double-edged: it improves long-form and hiYongshi Ye, Biao Fu, Chongxuan Huang, Yidong Chen, Xiaodong Shi
共 30 条热搜查看完整榜单