Peer Revenue Publish or profit

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computation and Language

cs.CL Watch Computation and Language

Saturday 9 May

Computation and Language · Cross-listed from cs.IR

The Interference Gap: Comparing Retrieval Bounds in Human Memory and RAG Systems

Dongxin Guo, Jikun Wu and Siu-Ming Yiu

How do retrieval bounds compare between human episodic memory and Retrieval-Augmented Generation (RAG) systems under semantic interference? We present a unified signal detection theory (SDT) framework that applies to both, and use it to fit behavioral and computational data in matched paradigms. Both systems show logarithmic accuracy decline with association count (fan), but humans exhibit lower interference sensitivity ($\alpha/\sigma = 0.41$) than dense passage retrieval ($\alpha/\sigma = 0.67$), with cognitively-inspired HippoRAG falling between the two ($\alpha/\sigma = 0.44$). Behavioral experiments ($N = 112$) and simulations validate the framework; parameter recovery confirms identifiability ($r \geq .93$) and model comparison favors the logarithmic specification over a power-law alternative ($\Delta$BIC $> 15$). We discuss encoding specificity, temporal context binding, and retrieval gating as candidate mechanisms whose causal role remains to be established. Six falsifiable predictions connect cognitive memory research with AI retrieval evaluation.
Read the full abstract

Computation and Language

Where Larger Models Excel: The Primacy of Constraint-Guided Reasoning

Guan-Yi Lin and Hen-Hsen Huang

Larger language models consistently outperform smaller ones on reasoning benchmarks, yet the reasoning differences underlying this gap remain underexplored. Across benchmarks in mathematics, physics, chemistry, and programming, we observe stable performance gaps: averaged over datasets, Qwen3-32B outperforms Qwen3-8B by 6.43%, while GPT-OSS-120B exceeds GPT-OSS-20B by 7.38%. To study the reasoning differences behind these gains, we develop AdvCluster, an automated framework that identifies questions where the larger model shows a stable advantage, extracts fine-grained advantage descriptions from paired reasoning traces produced by larger and smaller models, and organizes them through semantic clustering with quantitative evaluation and selection guided by a reviewer model. Our analysis yields a systematic taxonomy of larger model reasoning advantages, spanning both common advantages that recur across domains and specialized advantages associated with particular domains. Across these patterns, a recurring theme is Constraint-Guided Reasoning: larger models are better at identifying explicit and implicit constraints, organizing them into structured reasoning, and using them to rule out infeasible paths and verify intermediate steps.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Mechanistic Analysis of Alignment Algorithms in Language Models

Aarush Sinha, Ishan Garg, Veeraraju Elluru et al. (+2 more)

Post-training alignment algorithms are predominantly evaluated as black boxes, obscuring how they reshape language models' internal computations. We present a systematic mechanistic analysis of six preference-optimization methods: PPO, DPO, SimPO, ORPO, GRPO, and KTO across three open-weight model families. By integrating layer-wise linear probing, Sparse Autoencoders, and crosscoders, we localize preference representations and quantify alignment-induced geometric transformations in latent space. We find that preference signals consistently concentrate in early–mid or mid–late layers, but different objectives induce qualitatively distinct representational shifts. KTO and GRPO enhance linear separability through constructive feature sharing and sparse, high-salience recruitment. In contrast, DPO and ORPO degrade separability via non-constructive geometric rotation and feature attenuation, while PPO and SimPO largely preserve baseline geometry. These transformations exhibit architecture-dependent variability, demonstrating that behavioral alignment does not imply uniform internal restructuring. Our findings establish alignment as a heterogeneous intervention, motivate standardized feature-level auditing for safety and interpretability, and highlight the need for mechanism-aware optimization objectives.
Read the full abstract

All authors Aarush Sinha, Ishan Garg, Veeraraju Elluru, Arth Singh, Kushal Garg

Computation and Language · Cross-listed from cs.SD

WASIL: In-the-Wild Arabic Spoken Interactions with LLMs

Zien Sheikh Ali, Hamdy Mubarak, Soon-Gyo Jung et al. (+3 more)

Large Language Models (LLMs) voice assistants are commonly built as cascaded Automatic Speech recognition (ASR) to LLM systems, where recognition errors can distort user intent. Dislikes may also arise from ambiguous, out-of-domain, or non-request turns, making it hard to isolate ASR effects. We release WASIL (it denotes connection or linking in Arabic): in-the-wild Arabic spoken interaction prompts with audio, ASR hypotheses, assistant responses, and explicit like/dislike feedback (8,529 turns; 14.2% dislikes), plus a 2,000-turn test set covering Modern Standard Arabic (MSA) and four major dialects with their labels. We provide low-cost gold transcripts via multi-ASR agreement-guided post-editing and annotate answerability (answerable, ambiguous/needs-clarification, unsupported, not-a-request/noise) to separate intrinsic unanswerability from ASR-induced degradation. Finally, we describe scalable reference-free evaluation of responses from ASR vs. gold transcripts using multi-judge LLM scoring.
Read the full abstract

All authors Zien Sheikh Ali, Hamdy Mubarak, Soon-Gyo Jung, Hunzalah Hassan Bhatti, Firoj Alam, Shammur Absar Chowdhury

Computation and Language

Two Ways to De-Bias an LLM-as-a-Judge: A Continuous-Score Comparison of Hierarchical Bayesian Calibration and Neural-ODE Score Transport

Andrea Morandi

[Abridged] Using a Large Language Model (LLM) as an automatic rater (LLM-as-a-judge) is cheap but potentially biased: some judges run lenient, others strict, the middle of the scale gets compressed, and verbose answers may be over-rewarded. A common remedy is post-hoc calibration: leave the cheap judge in place and, on a modest set of paired anchors, fit a transformation from raw judge scores to an estimate of the human rating. We compare two correctors that take opposing views on how this mapping should be modeled: a parametric, small-anchor hierarchical Bayesian linear correction with per-score uncertainty, and a non-parametric Neural-ODE (FFJORD) score-transport flow. Both are run head-to-head on UltraFeedback fine-grained_score (1700 paired examples, 200 held out), with calibration split into three operational sub-questions: population-mean recovery, per-item accuracy, and distributional-shape match. The headline result is that the choice between methods is primarily a data-budget question. Both correctors close the raw $+0.71$-point mean offset to within $\pm 0.08$ of the GPT-4 reference, at 100 and at 1500 anchors. Past that, the methods swap roles. With 100 anchors, the linear corrector reconstructs the human-score distribution roughly twice as well by KL divergence (0.031 vs. 0.058) and ties the flow on MAE. With 1500 anchors the flow wins on every metric (MAE 0.320 vs. 0.359, Pearson 0.922 vs. 0.896, KL 0.026 vs. 0.037). The Bayesian linear corrector saturates well below 1500 anchors: residual $\tanh$-shaped non-linearity is, by construction, structure a linear correction cannot fit. The flow keeps improving as labels grow. We translate these findings into an explicit decision rule for production deployments.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Emergent Semantic Role Understanding in Language Models

Carla Griffiths and Mirco Musolesi

Understanding how linguistic structure emerges in language models is central to interpreting what these systems learn from data and how much supervision they truly require. In particular, semantic role understanding ("who did what to whom") is a core component of meaning representation, yet it remains unclear whether it arises from pre-training alone or depends on task-specific fine-tuning. We study whether semantic role understanding emerges during language model pre-training or requires task-specific fine-tuning. We freeze decoder-only transformers and train linear probes to extract semantic roles, using performance to infer whether role information is already encoded in pre-training or learned during adaptation. Across model scales, we find that frozen representations contain substantial semantic role information, with performance improving but not fully matching fine-tuned models. This indicates partial but incomplete emergence from pre-training alone. We show that semantic role structure emerges from language modeling objectives, but its internal implementation shifts toward more distributed representations as model scale increases.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Agentic MIP Research: Accelerated Constraint Handler Generation

Liding Xu, Yugeng Zhou and Sebastian Pokutta

Mixed-integer programming (MIP) research is both mathematically sophisticated and engineering-intensive: testing an algorithmic hypothesis within a branch-and-cut solver requires substantial implementation, debugging, tuning, and large-scale benchmarking. We propose an agentic MIP research framework that shortens this feedback loop by embedding LLM agents into a solver-aware harness for generating, verifying, and evaluating plugins for the open-source solver SCIP. Propagation methods play a central role in accelerating MIP solving by exploiting global constraints. We instantiate our framework on the semantic lifting of MIP formulations into global constraints and the automatic construction of propagation-only SCIP constraint handlers. On the MIPLIB 2017 benchmark set, the framework successfully recovers global constraint structures from constraint programming and generates executable constraint detectors and propagation-only constraint handlers. Furthermore, the framework naturally extends to in-context learning within a sandboxed environment, enabling agents not only to tune and debug generated constraint handlers on real instances, but also to explore global constraint patterns in MIP problems and discover novel propagation strategies not yet implemented in SCIP. This framework allows us to systematically distinguish meaningful algorithmic improvements from low-value or overly costly candidates: the novel propagation methods successfully solved five additional instances within the explored benchmark. Overall, this framework demonstrates that LLM agents can autonomously navigate the complex MIP research loop, paving the way for a more automated solver development process.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Open Ontologies: Tool-Augmented Ontology Engineering with Stable Matching Alignment

Fabio Rovai

We present Open Ontologies, an open-source ontology engineering system implemented in Rust that integrates LLM-driven construction with formal OWL reasoning and ontology alignment via the Model Context Protocol. Our primary finding is that stable 1-to-1 matching is the dominant factor in ontology alignment quality: on the OAEI Anatomy track, it achieves F1 = 0.832 (P = 0.963, R = 0.733), competitive with state-of-the-art systems and exceeding all in precision. Ablation across five weight configurations shows that signal weights are irrelevant when stable matching is applied (F1 varies by less than 0.004), while removing stable matching drops F1 to 0.728. On the Conference track, the same method achieves F1 = 0.438. On tool-augmented ontology interaction, we find a surprising result: an LLM reading a raw OWL file (F1 = 0.323) performs worse than the same LLM with no file at all (F1 = 0.431), while structured MCP tool access achieves F1 = 0.717. This demonstrates that tool structure provides a qualitatively different mode of access that the LLM cannot replicate by reading raw syntax. The system ships as a single binary under the MIT licence.
Read the full abstract

Computation and Language

WorldSpeech: A Multilingual Speech Corpus from Around the World

Antonis Asonitis, Luca A. Lanzendörfer, Frédéric Berdoz et al. (+1 more)

Automatic speech recognition (ASR) performs well for high-resource languages with abundant paired audio-transcript data, but its accuracy degrades sharply for most languages due to limited publicly available aligned data. To this end, we introduce WorldSpeech, a 24 kHz multilingual speech corpus comprising 65k hours of aligned audio-transcript data across 76 languages, collected from diverse public sources including parliamentary proceedings, international broadcasts, and public-domain audiobooks. For 37 languages, WorldSpeech provides more than 200 hours of aligned speech, with 28 exceeding 500 hours and 24 surpassing 1k hours. Fine-tuning existing ASR models on WorldSpeech results in an average relative Word-Error-Rate reduction of 63.5% across 11 typologically diverse languages.
Read the full abstract

All authors Antonis Asonitis, Luca A. Lanzendörfer, Frédéric Berdoz, Roger Wattenhofer

Computation and Language · Cross-listed from cs.LG

Sparse Layers are Critical to Scaling Looped Language Models

Ryan Lee, Jacob Biloki, Edward J. Hu et al. (+1 more)

Looped language models repeat a set of transformer layers through depth, reducing memory costs and providing natural early-exit points at loop boundaries. However, looped models do not scale as favorably as standard transformers with unique layers. We compare standard and Mixture-of-Experts (MoE) transformers, with and without looping, and find two main results. First, we find Looped-MoE models scale better than the standard baseline while dense looped models do not. We trace this to routing divergence between loops: in Looped-MoE models, different experts are activated on each pass through the same shared layers, recovering expressivity without additional parameters. Our second finding is that looped models have better compute-quality trade-offs with early exits than standard models. Because each loop ends with the same layers that produce the final output, loop boundaries are superior exit points, as confirmed by earlier output convergence at these points. In sum, we provide a clear direction for scaling looped models: a Looped-MoE model with early exits can not only beat standard transformers at scale, but also enable significant memory and inference savings with minimal degradation in quality.
Read the full abstract

All authors Ryan Lee, Jacob Biloki, Edward J. Hu, Jonathan May

Computation and Language

Lost in Translation? Exploring the Shift in Grammatical Gender from Latin to Occitan

Ahan Chatterjee, Matthias Schöffel, Matthias Aßenmacher et al. (+2 more)

The diachronic evolution from Latin to the Romance languages involved a restructuring of the grammatical gender system from a tripartite configuration (masculine, feminine, neuter) to a bipartite one (masculine, feminine) in most Romance languages. In this work, we introduce an interpretable deep learning framework to investigate this phenomenon at both lexical and contextual levels. First, we show that conventional tokenization strategies are insufficiently robust for this low-resource historical setting, and that our proposed tokenizer improves performance over these baselines. At the lexical level, we evaluate the contribution of morphological features to gender prediction. At the contextual level, we quantify the contributions of different part-of-speech categories to grammatical gender prediction. Together, these analyses characterize the distribution of gender information between the lemma and its sentential context. We make our codebase, datasets, and results publicly available at \href{https://github.com/ahan-2000/Lost-in-Translation-}{https://github.com/ahan-2000/Lost-in-Translation-}.
Read the full abstract

All authors Ahan Chatterjee, Matthias Schöffel, Matthias Aßenmacher, Marinus Wiedner, Esteban Garces Arias

Computation and Language

Meow-Omni 1: A Multimodal Large Language Model for Feline Ethology

Jucheng Hu, Zhangquan Chen, Yulin Chen et al. (+9 more)

Deciphering animal intent is a fundamental challenge in computational ethology, largely because of semantic aliasing, the phenomenon where identical external signals (e.g., a cat's purr) correspond to radically different internal states depending on physiological context. Existing Multimodal Large Language Models (MLLMs) are blind to high-frequency biological time-series data, restricting them to superficial behavioural pattern matching rather than genuine latent-state reasoning. To bridge this gap, we introduce Meow-Omni 1, the first open-source, quad-modal MLLM purpose-built for computational ethology. It natively fuses video, audio, and physiological time-series streams with textual reasoning. Through targeted architectural adaptation, we integrate specialized scientific encoders into a unified backbone and formalize intent inference via physiologically grounded cross-modal alignment. Evaluated on MeowBench, a novel, expert-verified quad-modal benchmark, Meow-Omni 1 achieves state-of-the-art intent-recognition accuracy (71.16%), substantially outperforming leading vision-language and omni-modal baselines. We release the complete open-source pipeline including model weights, training framework, and the Meow-10K dataset, to establish a scalable paradigm for inter-species intent understanding and to advance foundation models toward real-world veterinary diagnostics and wildlife conservation.
Read the full abstract

All authors Jucheng Hu, Zhangquan Chen, Yulin Chen, Chengjie Hong, Liang Zhou, Tairan Wang, Sifei Li, Giulio Zhu, Feng Zhou, Yiheng Zeng, Suorong Yang, Dongzhan Zhou

Computation and Language

From Traditional Taggers to LLMs: A Comparative Study of POS Tagging for Medieval Romance Languages

Matthias Schöffel and Esteban Garces Arias

Part-of-speech (POS) tagging for Medieval Romance languages remains challenging due to orthographic variation, morphological complexity, and limited annotated resources. This paper presents a systematic empirical evaluation of large language models (LLMs) for POS tagging across three medieval varieties: Medieval Occitan, Medieval Catalan, and Medieval French. We compare traditional rule-based and statistical taggers with modern open-source LLMs under zero-shot prompting, few-shot prompting, monolingual fine-tuning, and cross-lingual transfer learning settings. Experiments on historically grounded datasets show that LLM-based approaches consistently outperform traditional taggers, with fine-tuning and multilingual training yielding the largest improvements. In particular, cross-lingual transfer learning substantially benefits under-resourced varieties, while targeted bilingual training can outperform broader multilingual configurations for specific target languages. The results highlight the importance of linguistic proximity and dataset characteristics when designing transfer strategies for historical NLP. These findings provide empirical insights into the applicability of modern neural methods to medieval text processing and provide practical guidance for deploying LLM-based POS tagging pipelines in digital humanities research. All code, models, and processed datasets are released for reproducibility.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

A Communication-Theoretic Framework for LLM Agents: Cost-Aware Adaptive Reliability

Hamed Omidvar and Vahideh Akhlaghi

Agents built on large language models (LLMs) rely on a range of reliability techniques, including retry, majority voting, and self-consistency, that have been developed in parallel rather than within a common analytical framework. We observe that an LLM sampled at temperature $T$ is a discrete stochastic channel $p(y \mid x)$ in the sense of Shannon's coding theory, and use this identity as the entry point for such a framework grounded in communication theory. Each of these techniques is a special case of one of six classical reliability operators: diversity combining, hybrid retransmission, iterative generator-critic decoding, rateless sampling, structured redundant verification, and difficulty-adaptive routing. Within the framework we give two closed-form results: a noise-variance threshold above which uniform averaging beats quality-weighted averaging, and a contractivity criterion for generator-critic refinement, consistent with a contractive-to-divergent transition we observe between 3B- and 14B-parameter models. We further introduce a cost-aware semantic-nearest-neighbor router whose single Lagrangian knob traverses the quality-cost frontier without retraining. Across six channel configurations spanning local and cloud models on 69 hard tasks, no fixed model-technique-budget choice dominates, motivating per-task allocation. On a 300-item hard split of MMLU, GSM8K, and HumanEval, our router occupies the full empirical Pareto frontier: at matched quality, its normalized cost is ${\approx}56$% lower than the strongest fixed technique; at matched normalized cost, it improves quality by ${\approx}7$% ($26$% over single-shot decoding). These results argue for consolidating these reliability techniques into a single tunable layer informed by channel coding.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Personalized Alignment Revisited: The Necessity and Sufficiency of User Diversity

Enoch Hyunwook Kang

Personalized alignment aims to adapt large language models to heterogeneous user preferences, yet the precise theoretical conditions for its statistical efficiency have not been formally established. This paper characterizes the conditions under which personalized alignment achieves O(1) online regret and log(1/epsilon) offline sample complexity. We show that these optimal rates depend on a specific user-diversity condition: the population of user-specific heads must span the latent reward directions that can alter the optimal response. We prove that this condition is both necessary and sufficient. When it holds, simple greedy algorithms achieve benchmark efficiency; when it fails, every learner in a natural admissible class incurs at least logarithmic regret. Our results identify user diversity as the fundamental driver of personalized identifiability.
Read the full abstract

Computation and Language

Fin-Bias: Comprehensive Evaluation for LLM Decision-Making under human bias in Finance Domain

Xiaoyu Hu and Jinman Zhao

Large language models (LLMs) are increasingly deployed in financial contexts, raising critical concerns about reliability, alignment, and susceptibility to adversarial manipulation. While prior finance-related benchmarks assess LLMs' capabilities in stock trading, they are often restricted to small sample and fail to demonstrate LLM susceptibility to context with potential human bias. We introduce Fin-Bias (financial herding under long and uncertain financial context), a benchmark for evaluating LLM investment decision-making when faced with uncertainty and possible human-biased opinions. Fin-Bias includes 8868 long firm-specific analyst reports, including firm aspects summarized and analyzed by sophisticated analysts with investment ratings (Bullish/Neutral/Bearish) spanning from various industries. We present large language models with firm analyst reports with/without analyst investment ratings and even with 'fake' rating, to get investment ratings generated by LLMs. Our results reveal that LLMs tend to herd the explicit bias in context. We also develop a method to detect potential human opinions, which can encourage LLMs to think independently, some models even exceed human performance in predicting future stock return.
Read the full abstract

Computation and Language

GRC: Unifying Reasoning-Driven Generation, Retrieval and Compression

Zhongtao Miao, Qiyu Wu and Yoshimasa Tsuruoka

Text embedding and generative tasks are usually trained separately based on large language models (LLMs) nowadays. This causes a large amount of training cost and deployment effort. Context compression is also a challenging and pressing task, which is vital to reasoning-driven generation, and agentic tasks requiring long context and continual learning. In this paper, we explore how to unify reasoning-driven generation, reasoning-enhanced text representation and context compression tasks in one forward pass for LLMs. Through meta latent tokens and a unified generative, representative and compressive tuning approach, we propose a training framework named GRC that bridges the three tasks. The trained models can accomplish three objectives in a single forward pass while maintaining modular, LEGO-style flexibility during inference. This design greatly reduces the deployment effort for retrieval-augmented generation (RAG) and achieves efficient inference and three times data utilization during training. Furthermore, this framework design enables a new paradigm for text embedding: self-reason-latent embeds, and a new generation paradigm, latent memory-augmented generation, where compressed and internalized KV cache with O(1) length is used as the updatable memory. We also propose hybrid paged attention to speed up the inference of our models. Extensive experiments on reasoning-intensive retrieval benchmarks, generative tasks, document compression, latency evaluation, and RAG settings demonstrate the effectiveness of our method and may shed light on the truly unified model that can handle reasoning-driven generation, embedding and compression tasks seamlessly.
Read the full abstract

Computation and Language

Dynamic Meta-Metrics: Source-Sentence Conditioned Weighting for MT Evaluation

Luke Zhang, Justin Vasselli, Aditya Khan et al. (+2 more)

We propose Dynamic Meta-Metrics (DMM), a framework for machine translation evaluation that learns source-sentence conditioned combinations of existing metrics. Rather than relying on a single static ensemble or language-specific weighting, DMM adapts the metric combination based on properties of the source segment. We study hard conditioning, which fits an interpretable combiner per cluster, and an exploratory soft-conditioned extension whose weights vary continuously with source-cluster responsibilities. We evaluate DMM on the WMT Metrics Shared Task data across multiple language pairs using pairwise agreement measures at the system and segment levels. Across settings, MLP-based combinations outperform linear and Gaussian process-based ensembles, and introducing soft conditioning yields gains over linear models.
Read the full abstract

All authors Luke Zhang, Justin Vasselli, Aditya Khan, York Hay Ng, En-Shiun Annie Lee

Computation and Language

Character-Level Transformer for Tajik-Persian Transliteration with a Parallel Lexical Corpus

Mullosharaf K. Arabov

This study addresses automatic transliteration from Tajik (Cyrillic script) to Persian (Perso-Arabic script). We present a curated, lexicographically verified parallel corpus of 52,152 Tajik–Persian words and short phrases, compiled from printed dictionaries, encyclopedic sources, and manually verified online resources. To the best of our knowledge, this is one of the largest publicly available word-level corpora for Tajik–Persian transliteration. Using this corpus, we train a character-level sequence-to-sequence Transformer model and evaluate it using Character Error Rate (CER) and exact-match accuracy. The Transformer achieves a CER of 0.3216 and an exact-match accuracy of 0.3133, outperforming both dictionary-based rule-based and recurrent neural baselines. With beam search (k=3), performance further improves to CER 0.3182 and accuracy 0.3215. We describe the data collection and preprocessing pipeline, model architecture, and experimental protocol, and report a part-of-speech analysis showing performance differences across lexical categories. All preprocessing scripts, deterministic splits into training, validation, and test sets, and training configurations are released to support reproducibility and further research on Tajik and related Persian dialects. The corpus supports research in character-level transliteration, cross-script NLP, and lexicographic applications.
Read the full abstract

Computation and Language

Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs

Guijin Son, Seungone Kim, Catherine Arnett et al. (+73 more)

Following the recent achievement of gold-medal performance on the IMO by frontier LLMs, the community is searching for the next meaningful and challenging target for measuring LLM reasoning. Whereas olympiad-style problems measure step-by-step reasoning alone, research-level problems use such reasoning to advance the frontier of mathematical knowledge itself, emerging as a compelling alternative. Yet research-level math benchmarks remain scarce because such problems are difficult to source (e.g., Riemann Bench and FrontierMath-Tier 4 contain 25 and 50 problems, respectively). To support reliable evaluation of next-generation frontier models, we introduce Soohak, a 439-problem benchmark newly authored from scratch by 64 mathematicians. Soohak comprises two subsets. On the Challenge subset, frontier models including Gemini-3-Pro, GPT-5, and Claude-Opus-4.5 reach 30.4%, 26.4%, and 10.4% respectively, leaving substantial headroom, while leading open-weight models such as Qwen3-235B, GPT-OSS-120B, and Kimi-2.5 remain below 15%. Notably, beyond standard problem solving, Soohak introduces a refusal subset that probes a capability intrinsic to research mathematics: recognizing ill-posed problems and pausing rather than producing confident but unjustified answers. On this subset, no model exceeds 50%, identifying refusal as a new optimization target that current models do not directly address. To prevent contamination, the dataset will be publicly released in late 2026, with model evaluations available upon request in the interim.
Read the full abstract

All authors Guijin Son, Seungone Kim, Catherine Arnett, Hyunwoo Ko, Hyein Lee, Hyeonah Kang, Jiang Longxi, Jin Yun, JungYup Lee, Kyungmin Lee, Sam Yoosuk Kim, Sang Park, Seunghyeok Hong, SeungJae Lee, Seungyeop Yi, Shinae Shin, SunHye Bok, Sunyoung Shin, Yonghoon Ji, Youngtaek Kim, Hanearl Jung, Akari Asai, Graham Neubig, Sean Welleck, Youngjae Yu, Akshelin R, Alexander B. Ivanov, Boboev Muhammadjon, Chae Young Han, Christian Stump, Cooper R. Anderson, Dmitrii Karp, Dohyun Kwon, Dongryung Yi, DoYong Kwon, Duk-Soon Oh, Eunho Choi, Giovanni Resta, Greta Panova, Huiyun Noh, Hyungryul Baik, Hyungsun Bae, Inomov Mashrafdzhon, Jeewon Kim, Jeong-Rae Kim, Ji Eun Lee, Jiaqi Liu, Jieui Kang, Jimin Kim, Jon-Lark Kim, Joonyeong Won, Junseo Yoon, Junwoo Jo, Kibeom Kim, Kiwoon Kwon, Mario Kummer, Max Mercer, Min Hoon Kim, Minjun Kim, Nahyun Lee, Ng Ze-An, Nicolas Libedinsky, Rafał Marcin Łochowski, Raphaël Lachièze-Rey, Robert Auffarth, Ruichen Zhang, Sejin Park, Seonguk Seo, Shin Jaehoon, Sunatullo, Taewoong Eom, Yeachan Park, Yongseok Jang, Youchan Oh, Zhaoyang Wang, Zoltán Kovács

Computation and Language

Language-Conditioned Visual Grounding with CLIP Multilingual

J. de Curtò, Mauro Liz and I. de Zarzà

Multilingual vision-language models exhibit systematic performance gaps across languages, but the mechanism remains ambiguous: cross-language divergence could arise from the visual encoder, the text branch, or their interaction. We resolve this ambiguity through a dense multilingual CLIP probe in which the visual encoder is held identical across thirteen typologically diverse languages and only the XLM-RoBERTa text branch varies. We evaluate two CLIP architectures spanning a 7x visual-encoder scale gap (XLM-R base + ViT-B/32, ~87M visual parameters; XLM-R large + ViT-H/14, ~632M) on 11 concepts and 210 images, and quantify cross-language agreement via cluster-mask IoU, top-percentile IoU, and Spearman rank correlation against an English reference (n=2,310 paired observations per language). Three findings emerge. First, low-resource languages (Arabic, Basque, Luxembourgish) incur a structural penalty at both backbone scales (Wilcoxon HR>LR p<10^-300; cluster-mask IoU gap +0.114 at base, +0.143 at large), isolating the deficit to the text branch. Second, scaling the encoder 7x widens the gap for structural failure cases (Basque {\Delta}=-0.056, Luxembourgish {\Delta}=-0.076) while improving Arabic ({\Delta}=+0.033), separating corpus-coverage from tokeniser-fertility failures. Third, peak similarity is preserved across languages (mean ratio 0.94 at large scale) while cluster-mask IoU drops sharply, identifying spatial misalignment, not signal collapse, as the dominant failure mode. At 3.4-3.9 Wh per 1,000 queries, dense-CLIP grounding is competitive with high-throughput inference budgets, positioning it as a practical substrate for energy-aware multilingual deployment.
Read the full abstract

Computation and Language

Phase Transitions in Affective Meaning Divergence: The Hidden Drift Before the Break

Napassorn Litchiowong

One partner says "Fine" meaning "resolution"; the other hears "surrender." The word is shared; the affective uptake is not. We formalize this as affective meaning divergence (AMD), the total-variation distance between interlocutors' anchor-conditioned affect distributions. Building on speech-act theory, common-ground accumulation, and entropy-regularized game theory, we derive a logit best-response map whose dynamics undergo a saddle-node bifurcation: when $\beta\alpha > 4$, a monotone increase in AMD-driven load produces an abrupt, hysteretic collapse of repair coordination. On Conversations Gone Awry (CGA-Wiki; $N = 652$), derailing conversations exhibit critical-slowing-down (CSD) signatures across multiple levels: lexical divergence variance ($p < 0.001$, $d = 0.36$), AMD variance ($p = 0.001$, $d = 0.26$), and dialog-act repair variance ($p = 0.016$, $d = 0.20$), all significant after correction and stronger than toxicity and sentiment baselines. AMD provides a distinct temporal signature, with retrospectively measured variance peaking at the bifurcation point while toxicity variance peaks earlier, and is the only indicator grounded in the theoretical framework. Boundary-condition analysis on CGA-CMV ($N = 1,169$) yields mixed but directionally consistent evidence.
Read the full abstract

Computation and Language

Evaluating Pragmatic Reasoning in Large Language Models: Evidence from Scalar Diversity

Ye-eun Cho

Evaluating pragmatic reasoning in large language models (LLMs) remains challenging because model behavior can vary depending on evaluation methods. Previous studies suggest that prompt-based judgments may diverge from models' internal probability distributions, raising questions about whether observed performance reflects underlying competence or task-induced behavior. This study examines this issue using scalar diversity as a graded diagnostic for pragmatic inference. Following Hu & Levy (2023), this study compares direct probability measurement and metalinguistic prompting across multiple models and experimental settings. The results show that neither evaluation method consistently outperforms the other and that pragmatic behavior varies substantially across model families, prompting strategies, and task structures. Moreover, scalar diversity gradients emerge only in specific model-condition combinations, suggesting that pragmatic reasoning in LLMs reflects an interaction between internal probabilistic representations and task-induced prompting behavior rather than a stable competence captured by a single evaluation paradigm. These findings highlight the central role of evaluation design in interpreting pragmatic abilities in LLMs.
Read the full abstract

Computation and Language

BiAxisBias: Evaluating LLM Bias Beyond a Single Prompt and a Single Explanation

Jialing Gan, Junhao Dong and Songze Li

LLM bias scores can depend on audit design. We introduce BiAxisBias, a prespecified audit varying task, role, perspective, sentiment, and wording over 200 stereotype statements while retaining forced Selection and Rationale as separate protocol readouts. Its main matrix spans eight LLMs and 401 templates (641,600 responses). Across five equivalent questions, 17.1% of 1,600 model-statement pairs change Selection. With three observations per unit in both arms, instability averages 10.5% across all ten three-wording subsets, versus 6.3% for three identical calls. Across four controlled task paradigms, 9/28 model pairs reverse; a seven-model factorial sensitivity identifies task-by-sentiment as the largest two-way component (raw eta-squared = 0.0465). In 10,000 equal-budget resampling draws, mean absolute error against a declared 18-condition finite reference is 10.63 points for one-template concentration, 1.86 for matrix-wide simple random sampling, and 1.75 for condition stratification; ranking inversions are 20.6%, 4.8%, and 5.5%. Thus broad coverage drives the gain, while stratification has only a modest score-error advantage and no ranking advantage over random sampling. In a separate forced-output diagnostic, task-specific Selection mappings and judge-coded Rationale stance disagree in 34.2% of 7,959 dual-valid responses (31.2% versus 3.0% by direction). This diagnoses output-contract sensitivity, not two validated measures of one construct.
Read the full abstract

Computation and Language

A Quantum Inspired Variational Kernel and Explainable AI Framework for Cross Region Solar and Wind Energy Forecasting

Pavan Manjunath and Thomas Prufer

Reliable short horizon forecasting of solar and wind generation is a structural prerequisite of any modern power system yet most published forecasters are tuned and evaluated on a single climatic regime and most algorithmic novelty has been concentrated either on classical recurrent networks or on monolithic foundation models that combine forecasting and explanation We develop a four stage hybrid framework that separates these concerns The first stage acquires hourly generation irradiance and surface weather records through public application programming interfaces The second stage trains three classical baselines autoregressive integrated moving average gradient boosted regression trees and a two layer long short term memory network and produces a strong point forecast together with a residual error series The third stage corrects the residual through a quantum inspired variational kernel built on a six qubit hardware efficient ansatz with three repeated entangling layers The fourth stage uses generative artificial intelligence strictly as an explainability layer that reads the measured benchmark numbers and produces a structured natural language interpretation Across three regions drawn from open public archives Iberian solar North Sea wind and a mixed Texas trace the proposed configuration stays within one percentage point of the strongest classical baseline on the in domain forecasting task and the quantum inspired kernel separates calm and stormy weather regimes with a Fisher discriminant ratio approximately fifteen fold higher than a tuned radial basis kernel
Read the full abstract

Computation and Language

Agent Collectives Should Not Detect Their Own Imposters: A Chess Case Study

Alexandre Le Mercier, Chris Develder and Thomas Demeester

A collective of AI agents collaborating on a task has the potential to outclass any individual agent for that task. We study the robustness of such collectives against possible imposters, i.e., agents that deliberately try to mislead their peers. Since a single imposter could undo the collective's advantage, we need to detect them. We consider two strategies: (i) incorporate imposter detection into the participating agents, or (ii) use a dedicated imposter detector outside the collective. We investigate this empirically on Gambit, a testbed in which 4 reasoning agents deliberate on chess moves. The setting is small but still challenging for frontier models. Chess allows objective, quantitative assessment (via a state-of-the-art chess engine) of both the gain of using a collective and the damage done by imposters. We find that merely warning the agents of potential imposter presence is not beneficial: it degrades decisions when no imposter is present, provokes reactions ranging from self-accusation to scapegoating, inflates token use, and reveals to the imposter how it was uncovered. We therefore recommend a detector that reads the collective's deliberation but never joins it and only returns a verdict. Such a detector must recalibrate to new attack strategies after very few examples, rather than wait for full retraining. In our benchmark, a 3B language model with a meta-trained classification head achieves that: a single gradient step on 20 labeled examples suffices to adapt to an unseen imposter strategy. At matched zero-shot accuracy, this detector yields 8x the adaptation gain of standard finetuning, at 14x lower training cost. We release the Gambit benchmark, with 37,352 labeled deliberations spanning 240 evolved imposter strategies. Code and data: https://anonymous.4open.science/r/gambit.
Read the full abstract

Computation and Language

LLiMba: Sardinian on a Single GPU – Adapting a 3B Language Model to a Vanishing Romance Language

Luca Ballore

Sardinian, a Romance language with roughly one million speakers, has minimal presence in modern NLP. Commercial services do not support it, and current language models do not produce it reliably. We present LLiMba, a 3B parameter Sardinian-ready model adapted from Qwen2.5-3B-Instruct through continued pretraining (CPT) and supervised fine-tuning (SFT) on a single 24 GB consumer GPU. The corpus contains 11.5 million tokens of Sardinian spanning LSC, Logudorese, and Campidanese, augmented with 2.4 million tokens of related Romance text as replay against register blurring. After CPT the model reaches a perplexity of 6.76 on held out Sardinian and outperforms the base across all six FLORES-200 directions. We compare five SFT configurations under matched conditions: full fine-tuning, LoRA r64, rsLoRA r128, rsLoRA r256, and DoRA r256. rsLoRA r256 wins on every direction into Sardinian, reaching 28.5 BLEU from English against 17.3 after CPT and 21.0 with full fine-tuning. The rank ablation places r128 between LoRA r64 and rsLoRA r256 on BLEU but reveals failure modes invisible to the metric, including leakage across scripts no other variant produces. LoRA r64 retains less factual content from SFT than configurations at higher rank and produces more confident fabrications, though all methods fabricate on content absent from training. DoRA r256 yields the smallest gap between training and evaluation but the worst factual accuracy. The findings indicate that adapter capacity matters more than the choice among LoRA variants for adapting a Romance pretrained base to a low resource Romance target, that stronger regularization is not uniformly beneficial, and that translation metrics smoothly order configurations whose qualitative behavior differs categorically. Perplexity comparisons across scripts must account for byte fallback tokenization, which deflates the metric for scripts other than Latin.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Relative Kinetic Utility: Calibrating Cross-Layer Credit for Global Structured LLM Pruning

Tianhao Qian, Guilin Qi and Jiayu Chen

Global structured pruning requires channels from different layers to compete under a shared sparsity budget, raising two coupled challenges: identifying which channels should be retained and making their scores comparable across layers. Raw channel scores can contain block-common scale that leaves within-block ordering unchanged but distorts model-wide competition. Our experiment indicates that similar layer-wise allocations can retain substantially different FFN channels, so layer allocation alone does not determine channel identity. Motivated by this separation, we introduce Global Relative Kinetic Utility (Global RKU), a label-free criterion that separates channel importance estimation from cross-layer comparison. Global RKU measures channel participation using a final-hidden-state activation-gradient signal, then applies block-relative normalization to mitigate block-common scale while preserving within-block ordering, requires only unlabeled calibration inputs, and produces a static pruning topology in a single calibration stage. Under questions-only calibration on Qwen-2.5-7B, RKU-GISP Mean3 margins are -0.98, +3.79, and +8.61 points at 30%, 40%, and 50% sparsity, respectively (average +3.81). Additional Qwen evaluations cover non-mathematical reasoning, recovery, held-out transfer, and physical deployment. Separately, replacing Wiki16K with questions-only Q16K improves RKU's Mean3 at every tested sparsity on Qwen, Llama, and Gemma. Our ablation study shows relative-normalization gains of 14.42 and 5.53 Mean3 points at 40% and 50% sparsity, respectively; the common-seed audit is positive in all 27 seed-task comparisons.
Read the full abstract

Computation and Language

Dolphin-CN-Dialect: Where Chinese Dialects Matter

Yangyang Meng, Huihang Zhong, Guodong Lin et al. (+6 more)

We present Dolphin-CN-Dialect, a streaming-capable ASR model with a focus on Chinese and dialect-rich scenarios. Compared to the previous version, Dolphin-CN-Dialect introduces substantial improvements in data processing, tokenization, training stability, and data sampling strategies. To address the challenges of highly imbalanced dialect data, we propose a temperature-based sampling strategy that effectively balances standard Mandarin and low-resource dialects, leading to significant gains in dialect recognition performance. In addition, we redesign the tokenizer to better align with linguistic characteristics, adopting character-level modeling for Chinese and subword modeling for English, while introducing extensible dialect tokens. Experimental results show that Dolphin-CN-Dialect achieves improvement in dialect recognition accuracy and CER reduction compared to Dolphin. Furthermore, Dolphin-CN-Dialect reaches competitive performance with recent SOTA open-source ASR models, while maintaining a significantly smaller model size. Dolphin-CN-Dialect supports both streaming and non-streaming inference, enabling a practical balance between latency and accuracy. It also provides flexible customization through hotword support and efficient deployment optimized for specialized hardware. These improvements make Dolphin-CN-Dialect a strong and practical solution for real-world multi-dialect ASR applications.
Read the full abstract

All authors Yangyang Meng, Huihang Zhong, Guodong Lin, Guanbo Wang, Hu Du, Zhiming Shao, Yukai Huang, Ke Li, Wei-Qiang Zhang

Computation and Language

Improving Lexical Difficulty Prediction with Context-Aligned Contrastive Learning and Ridge Ensembling

Wicaksono Leksono Muhamad, Joanito Agili Lopo, Tsamarah Rana Nugraha et al. (+2 more)

Lexical difficulty prediction is a fundamental problem in language learning and readability assessment, requiring models to estimate word difficulty across different first-language (L1) backgrounds. However, existing approaches rely on regression-only training with scalar supervision, which does not explicitly structure the representation space, limiting their ability to capture cross-lingual alignment and ordinal difficulty. To mitigate these issues, we propose Context-Aligned Contrastive Regression, which integrates Ridge regression ensemble with two complementary objectives, i.e., Cross-View Context and Ordinal Soft Contrastive Learning. Experiments on three L1 datasets show that (i) contrastive objectives improve cross-lingual representation alignment while preserving language-specific nuances, (ii) the learned representations capture the ordinal structure of lexical difficulty, and (iii) the ensemble effectively mitigates systematic biases of individual models, leading to more stable performance across difficulty levels.
Read the full abstract

All authors Wicaksono Leksono Muhamad, Joanito Agili Lopo, Tsamarah Rana Nugraha, Ahmad Cahyono Adi, Muhammad Oriza Nurfajri