Computation and Language · Cross-listed from cs.LG
State Stream Transformer (SST) V2: Parallel Training of Nonlinear Recurrence for Latent Space Reasoning
Thea Aviss
Current
transformers discard their rich latent residual stream between positions, reconstructing latent
reasoning context at each new position and leaving potential
reasoning capacity untapped. The State Stream
Transformer (SST) V2 enables parameter-efficient
reasoning in continuous latent space through an FFN-driven nonlinear recurrence at each decoder layer, where latent states are streamed horizontally across the full sequence via a learned blend. This same mechanism supports continuous latent deliberation per position at inference time, dedicating additional FLOPs to exploring abstract
reasoning before committing to a
token. A two-pass parallel training procedure approximates the sequential recurrence, making co-training computationally practical. Hidden state analysis shows that the state stream facilitates
reasoning through sharp, content-dependent reorganisations in continuous latent space; the LM head exposes the resulting latent belief states through the output distribution, while the state stream carries them forward to influence future positions. A learned probe shows that at the first generated
token, the latent state already predicts whether the eventual answer will survive or break under additional latent computation for every subsequent position. Co-trained into an existing 27B backbone using only a small
dataset of GSM8K examples and evaluated using an oracle to route each question to a depth of one to four recurrent forward passes per
token, the SST achieves an architectural capacity bound of 61.11% on out-of-distribution GPQA-Diamond, a +15.15 point gain over a fine-tuning-matched baseline, and cuts that same baseline's remaining GSM8K errors by 46%. Together, these results provide a method for efficiently training a nonlinear recurrence and show that the state stream provides additional
reasoning capacity beyond that of an otherwise training-matched
transformer.
Read the full abstract ▾
Computation and Language
Confidence Estimation in Automatic Short Answer Grading with LLMs
Longwei Cong, Sonja Hahn, Sebastian Gombert et al. (+3 more)
Automatic Short Answer Grading (ASAG) with generative large
language models (LLMs) has recently demonstrated strong performance without task-specific fine-tuning, while also enabling the generation of synthetic feedback for educational assessment. Despite these advances, LLM-based grading remains imperfect, making reliable confidence estimates essential for safe and effective human-AI collaboration in educational decision-making. In this work, we investigate confidence estimation for ASAG with LLMs by jointly considering model-based confidence signals and
dataset-derived uncertainty. We systematically compare three model-based confidence estimation strategies, namely verbalizing, latent, and consistency-based confidence estimation, and show that model-based confidence alone is insufficient to reliably capture uncertainty in ASAG. To address this limitation, we propose a hybrid confidence framework that integrates model-based confidence signals with an explicit estimate of
dataset-derived aleatoric uncertainty. Aleatoric uncertainty is operationalized by clustering semantically embedded student responses and quantifying within-cluster heterogeneity. Our results demonstrate that the proposed hybrid confidence measure yields more reliable confidence estimates and improves selective grading performance compared to single-source approaches. Overall, this work advances confidence-aware LLM-based grading for human-in-the-loop assessment, supporting more trustworthy AI-assisted educational assessment systems.
Read the full abstract ▾
All authors Longwei Cong, Sonja Hahn, Sebastian Gombert, Leon Camus, Hendrik Drachsler, Ulf Kroehne
Computation and Language
RSAT: Structured Attribution Makes Small Language Models Faithful Table Reasoners
Jugal Gajjar and Kamalasankari Subramaniakuppusamy
When a
language model answers a table question, users have no way to verify which
cells informed which
reasoning steps. We introduce RSAT, a method that trains small
language models (SLMs, 1-8B) to produce step-by-step
reasoning with
cell-level citations grounded in table evidence. Phase 1 (SFT) teaches a structured JSON output format from verified
reasoning traces. Phase 2 (GRPO) optimizes a composite reward centered on NLI-based faithfulness, alongside citation validity and parsimony. Across six models from two families-Qwen 2.5 (1.5B/3B/7B) and Llama 3 (1B/3B/8B)-RSAT improves faithfulness 3.7$\times$ over SFT alone (0.224$\rightarrow$0.826), with near-perfect citation validity (0.992). Post-hoc attribution collapses below 13% format success, confirming that attribution must be integrated into
reasoning, not retrofitted. Ablations show the faithfulness reward is essential: removing it drops faithfulness from 0.97 to 0.03.
Read the full abstract ▾
Computation and Language · Cross-listed from cs.NI
RouteProfile: Graph-Based Profiling for Cold-Start LLM Routing
Jingjun Xu, Hongji Pu, Tao Feng et al. (+3 more)
LLM routing is increasingly important for selecting suitable models under diverse user needs and deployment constraints, but its practical effectiveness depends on continual adaptation to emerging queries and newly released models. New-LLM integration is particularly challenging, as newly released models lack the query-response-reward interactions required for router training and cannot be profiled as directly as new queries via semantic
embeddings. Existing profiles are limited: LLM-generated descriptions are often coarse, while interaction-based
embeddings are costly to construct. To address this problem, we propose RouteProfile, a graph-based profiling framework that constructs LLM profiles from public signals in technical reports or model cards, including model family, model description, reported
benchmark scores, and
benchmark domains. RouteProfile organizes these heterogeneous signals into a graph and studies profile construction along four dimensions: organizational form, representation type, aggregation depth, and learning configuration. We evaluate RouteProfile in training-free cold-start routing and new-LLM integration settings. Experiments show that: (1) structured profiles outperform flat baselines in training-free cold-start routing; (2) model family metadata is more reliable than
benchmark domain information; and (3) effective new-LLM integration requires profile-router co-design. Overall, our findings highlight the importance of profile design for enabling routing systems to adapt to the evolving model ecosystem.
Read the full abstract ▾
All authors Jingjun Xu, Hongji Pu, Tao Feng, Haozhen Zhang, Jiaxuan You, Ge Liu
Computation and Language · Cross-listed from cs.LG
Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback
Yikai Wang, Shang Liu and Jose Blanchet
Reinforcement learning from human feedback (RLHF) is a central post-training tool for aligning large
language models, but its training reward is only a learned proxy for true human utility. This creates a decision problem under objective misspecification: the policy is optimized against an estimated reward, while deployment performance is governed by an unobserved population preference. The resulting gap leads to reward over-optimization, where proxy reward keeps improving after true quality deteriorates. We propose distributionally robust regret optimization (DRRO) for RLHF with a Wasserstein ambiguity set over reward laws, using promptwise $\ell_p$ distances between reward vectors as transport costs. Unlike standard distributionally robust optimization, which pessimizes worst-case value, DRRO pessimizes worst-case regret relative to the best policy under the same plausible reward perturbation. We show that the expressive-policy problem decomposes into promptwise regret problems. For each
prompt, the inner adversary has a dual-norm closed form; under the $\ell_1$ transport cost used by our
algorithm, the optimizer has a water-filling structure. These results lead to a practical policy-gradient
algorithm that adds a simple sampled bonus to GRPO-style training. Theory and experiments both show that DRRO is less over-pessimistic than standard DRO and mitigates over-optimization more effectively than existing baselines.
Read the full abstract ▾
Computation and Language
Timing is Everything: Temporal Scaffolding of Semantic Surprise in Humor
Yuxi Ma, Yongqian Peng, Junchen Lyu et al. (+2 more)
Humor is a fundamental cognitive phenomenon in which humans derive pleasure from the expectation violations and their resolution, exemplifying the
brain's dynamic capacity for predictive processing. Classical humor theories emphasize semantic incongruity as the primary driver of amusement, yet overlook temporal dynamics despite comedians' intuition that "timing is everything." The extent to which temporal structure contributes to humor appreciation and how it interacts with semantic content remains poorly understood. Here, we propose the Dual Prediction Violation (DPV) framework to capture the interplay between content and timing. By analyzing 828 professional Chinese stand-up performances, we show that temporal features substantially outweigh semantic incongruity in predicting audience appreciation. Specifically, we find that peak semantic violations matter more than average incongruity levels, and pauses systematically lengthen before high-surprise punchlines–a strategic coupling that distinguishes successful from unsuccessful performances. These findings reframe humor as temporally scaffolded, where timing and semantic content operate in strategic coordination rather than independently. Our DPV framework bridges humor theory with predictive processing, demonstrating that temporal structure plays a central role in naturalistic humor appreciation with implications for understanding multi-scale prediction integration in linguistic processing.
Read the full abstract ▾
All authors Yuxi Ma, Yongqian Peng, Junchen Lyu, Chi Zhang, Yixin Zhu
Computation and Language · Cross-listed from cs.LG
Technical Report: Activation Residual Hessian Quantization (ARHQ) for Low-Bit LLM Quantization
YiFeng Wang, Zhun Sun and Keisuke Sakaguchi
We present Activation Residual Hessian Quantization (ARHQ), a post-training weight splitting method designed to mitigate error propagation in low-bit activation-weight quantization. By constructing an input-side residual Hessian from activation quantization residuals (G_x), ARHQ analytically identifies and isolates error-sensitive weight directions into a high-precision low-rank branch. This is achieved via a closed-form truncated SVD on the scaled weight
matrix W G^{1/2}_x . Experimental results on Qwen3-4B-Thinking-2507 demonstrate that ARHQ significantly improves layer-wise SNR and preserves downstream
reasoning performance on ZebraLogic even under aggressive quantization. The code is available at https://github.com/BeautMoonQ/ARHQ.
Read the full abstract ▾
Computation and Language
Cultural Benchmarking of LLMs in Standard and Dialectal Arabic Dialogues
Muhammad Dehan Al Kautsar, Saeed Almheiri, Momina Ahsan et al. (+13 more)
There is a significant gap in evaluating cultural
reasoning in LLMs using conversational
datasets that capture culturally rich and dialectal contexts. Most Arabic
benchmarks focus on short text snippets in Modern Standard Arabic (MSA), overlooking the cultural nuances that naturally arise in dialogues. To address this gap, we introduce ArabCulture-Dialogue, a culturally grounded conversational
dataset covering 13 Arabic-speaking countries, in both MSA and each country's respective dialect, spanning 12 daily-life topics and 54 fine-grained subtopics. We utilize the
dataset to form three benchmarking tasks: (i) multiple-choice cultural
reasoning, (ii) machine translation between MSA and dialects, and (iii) dialect-steering generation. Our experiments indicate that the performance gap between MSA and Arabic dialects still exists, whereby the models perform worse on all three tasks in the dialectal setup, compared to the MSA one.
Read the full abstract ▾
All authors Muhammad Dehan Al Kautsar, Saeed Almheiri, Momina Ahsan, Bilal Elbouardi, Younes Samih, Sarfraz Ahmad, Amr Keleg, Omar El Herraoui, Kareem Elzeky, Abed Alhakim Freihat, Mohamed Anwar, Zhuohan Xie, Junhong Liang, Mohammad Rustom Al Nasar, Preslav Nakov, Fajri Koto
Computation and Language
ViLegalNLI: Natural Language Inference for Vietnamese Legal Texts
Nhung Thi-Hong Duong, Mai Ngoc Ho, Tin Van Huynh et al. (+1 more)
In this article, we introduce ViLegalNLI, the first large-scale Vietnamese Natural
Language Inference (NLI)
dataset specifically constructed for the legal domain. The
dataset consists of 42,012 premise-hypothesis pairs derived from official statutory documents and annotated with binary inference labels (Entailment and Non-entailment). It covers multiple legal domains and reflects realistic legal
reasoning scenarios characterized by structured logic, conditional clauses, and domain-specific terminology. To construct ViLegalNLI, we propose a semi-automatic data generation framework that integrates large
language models for controlled hypothesis generation and systematic quality validation procedures. The framework incorporates artifact mitigation strategies and cross-model validation to improve annotation reliability and ensure legal consistency. The resulting
dataset captures diverse
reasoning patterns, including paraphrasing, logical implication, and legally invalid inferences, thereby providing a comprehensive
benchmark for Vietnamese legal inference tasks. We conduct extensive experiments on the ViLegalNLI using
multilingual models, Vietnamese-specific pretrained
language models, and instruction-tuned large
language models. The results show that few-shot LLM configurations consistently achieve superior performance, while performance is significantly influenced by hypothesis length, lexical overlap, and
reasoning complexity. Cross-domain evaluations further reveal the challenges of generalizing legal inference across distinct legal fields. Overall, ViLegalNLI establishes a foundational
benchmark for Vietnamese legal NLI and supports future research in legal
reasoning, statutory text understanding, and the development of reliable AI systems for legal analysis and decision support. The
dataset is publicly available for research purposes.
Read the full abstract ▾
All authors Nhung Thi-Hong Duong, Mai Ngoc Ho, Tin Van Huynh, Kiet Van Nguyen
Computation and Language
How Frontier LLMs Adapt to Neurodivergence Context: A Measurement Framework for Surface vs. Structural Change in System-Prompted Responses
Ishan Gupta and Pavlo Buryi
We examine if frontier chat-based large
language models (LLMs) adjust their outputs based on neurodivergence (ND) context in system
prompts and describe the nature of these adjustments. Specifically, we propose NDBench, a 576-output
benchmark involving two frontier models, three system
prompt types (baseline, ND-profile assertion, and ND-profile assertion with explicit instructions for adjustments), four canonical ND profiles, and 24
prompts across four categories, one of which involves an
adversarial masking strategy. Four trends emerge consistently from our findings. First, LLMs show significant adaptation under ND context, where fully instructed conditions yield lengthier and more structured outputs, characterized by higher
token counts, more headings, and more granular steps (p < 10^-8, Holm-corrected). Second, such adaptation is largely structural in nature: although list density does not change much, there is a marked rise in the frequency of headings and per-step detail. Third, ND persona assertion alone fails to suppress potentially harmful tendencies, as masking-reinforcement decreases only in explicitly instructed cases (36-44% reduction); the reduction rate barely changes in persona assertion conditions. Moreover, reliability analysis of LLM-based harm assessment reveals that only two out of the six dimensions (masking and reinforcement, validation quality) exceed the pre-defined inter-judge agreement criterion (alpha >= 0.67) and thus can be considered primary results. NDBench is made publicly available along with its
prompts, outputs, code, and other resources, forming a reproducible framework for
auditing future LLMs' adaptation to ND awareness.
Read the full abstract ▾
Computation and Language
NorBERTo: A ModernBERT Model Trained for Portuguese with 331 Billion Tokens Corpus
Enzo S. N. Silva, Pablo B. Costa, Raphael C. Vlasman et al. (+8 more)
High-quality corpora are essential for advancing Natural
Language Processing (NLP) in Portuguese. Building on previous encoder-only models such as BERTimbau and Albertina PT-BR, we introduce NorBERTo, a modern encoder based on the ModernBERT architecture, featuring long-context support and efficient
attention mechanisms. NorBERTo is trained on Aurora-PT, a newly curated Brazilian Portuguese corpus comprising 331 billion GPT-2
tokens collected from diverse web sources and existing
multilingual datasets. We systematically
benchmark NorBERTo against Strong baselines on semantic similarity, textual entailment and classification tasks using standardized
datasets such as ASSIN 2 and PLUE. On PLUE, NorBERTo-large achieves the best results among the encoder models we evaluated, notably reaching 0.9191 F1 on MRPC and 0.7689 accuracy on RTE. On ASSIN 2, NorBERTo-large attains the highest entailment F1 (~0.904) among all encoders considered, although Albertina-900M and BERTimbau-large still hold an advantage. To the best of our knowledge, Aurora-PT is currently the largest openly available monolingual Portuguese corpus, surpassing previous resources. NorBERTo provides a modern, mid-sized encoder designed for realistic deployment scenarios: it is straight-forward to fine-tune, efficient to serve, and well suited as a backbone for
retrieval-augmented generation and other downstream Portuguese NLP systems.
Read the full abstract ▾
All authors Enzo S. N. Silva, Pablo B. Costa, Raphael C. Vlasman, Rosimeire P. Costa, Henrique L. P. Silva, Lucas F. A. O. Pellicer, Guilherme Rinaldo, Renato A. Almeida, Darian S. R. Rabbani, Cinthya O. Oestreich, Vinicius F. Caridá
Computation and Language · Cross-listed from cs.LG
Exploration Hacking: Can LLMs Learn to Resist RL Training?
Eyon Jang, Damon Falck, Joschka Braun et al. (+6 more)
Reinforcement learning (RL) has become essential to the post-training of large
language models (LLMs) for
reasoning, agentic capabilities and
alignment. Successful RL relies on sufficient exploration of diverse actions by the model during training, which creates a potential failure mode: a model could strategically alter its exploration during training to influence the subsequent training outcome. In this paper we study this behavior, called exploration hacking. First, we create model organisms of selective RL resistance by fine-tuning LLMs to follow specific underperformance strategies; these models can successfully resist our RL-based capability elicitation in agentic biosecurity and AI R&D environments while maintaining performance on related tasks. We then use our model organisms to evaluate detection and mitigation strategies, including monitoring, weight noising, and SFT-based elicitation. Finally, we show that current frontier models can exhibit explicit
reasoning about suppressing their exploration when provided with sufficient information about their training context, with higher rates when this information is acquired indirectly through the environment. Together, our results suggest exploration hacking is a possible failure mode of RL on sufficiently capable LLMs.
Read the full abstract ▾
All authors Eyon Jang, Damon Falck, Joschka Braun, Nathalie Kirch, Achu Menon, Perusha Moodley, Scott Emmons, Roland S. Zimmermann, David Lindner
Computation and Language · Cross-listed from cs.AI
Synthetic Computers at Scale for Long-Horizon Productivity Simulation
Tao Ge, Baolin Peng, Hao Cheng et al. (+1 more)
Realistic long-horizon productivity work is strongly conditioned on user-specific computer environments, where much of the work context is stored and organized through directory structures and content-rich artifacts. To scale synthetic data creation for such productivity scenarios, we introduce Synthetic Computers at Scale, a scalable methodology for creating such environments with realistic folder hierarchies and content-rich artifacts (e.g., documents, spreadsheets, and presentations). Conditioned on each synthetic computer, we run long-horizon simulations: one
agent creates productivity objectives that are specific to the computer's user and require multiple professional deliverables and about a month of human work; another
agent then acts as that user and keeps working across the computer – for example, navigating the filesystem for grounding, coordinating with simulated collaborators, and producing professional artifacts – until these objectives are completed. In preliminary experiments, we create 1,000 synthetic computers and run long-horizon simulations on them; each run requires over 8 hours of
agent runtime and spans more than 2,000 turns on average. These simulations produce rich experiential learning signals, whose effectiveness is validated by significant improvements in
agent performance on both in-domain and out-of-domain productivity evaluations. Given that personas are abundant at billion scale, this methodology can in principle scale to millions or even billions of synthetic user worlds with sufficient compute, enabling broader coverage of diverse professions, roles, contexts, environments, and productivity needs. We argue that scalable synthetic computer creation, together with at-scale simulations, is highly promising as a foundational substrate for
agent self-improvement and agentic reinforcement learning in long-horizon productivity scenarios.
Read the full abstract ▾
All authors Tao Ge, Baolin Peng, Hao Cheng, Jianfeng Gao
Computation and Language
On the Proper Treatment of Units in Surprisal Theory
Samuel Kiegeland, Vésteinn Snæbjarnarson, Tim Vieira et al. (+1 more)
Surprisal theory links human processing effort to the predictability of an upcoming linguistic unit, but empirical work often leaves the notion of a unit underspecified. In practice, experimental stimuli are segmented into linguistically motivated units (e.g., words), while pretrained
language models assign probability mass to a fixed
token alphabet that typically does not align with those units. As a result, surprisal-based predictors depend implicitly on ad hoc procedures that conflate two distinct modeling choices: the definition of the unit of analysis and the choice of regions of interest over which predictions are evaluated. In this paper, we disentangle these choices and give a unified framework for
reasoning about surprisal over arbitrary unit inventories. We argue that surprisal-based analyses should make these choices explicit and treat tokenization as an implementation detail rather than a scientific primitive.
Read the full abstract ▾
All authors Samuel Kiegeland, Vésteinn Snæbjarnarson, Tim Vieira, Ryan Cotterell
Computation and Language · Cross-listed from cs.CV
Sudong Wang, Weiquan Huang, Xiaomin Yu et al. (+9 more)
The standard post-training recipe for large
multimodal models (LMMs) applies supervised fine-tuning (SFT) on curated demonstrations followed by reinforcement learning with verifiable rewards (RLVR). However, SFT introduces distributional drift that neither preserves the model's original capabilities nor faithfully matches the supervision distribution. This problem is further amplified in
multimodal reasoning, where perception errors and
reasoning failures follow distinct drift patterns that compound during subsequent RL. We introduce PRISM, a three-stage pipeline that mitigates this drift by inserting an explicit distribution-
alignment stage between SFT and RLVR. Building on the principle of on-policy
distillation (OPD), PRISM casts
alignment as a black-box, response-level
adversarial game between the policy and a Mixture-of-Experts (MoE) discriminator with dedicated perception and
reasoning experts, providing disentangled corrective signals that steer the policy toward the supervision distribution without requiring access to teacher logits. While 1.26M public demonstrations suffice for broad SFT initialization, distribution
alignment demands higher-fidelity supervision; we therefore curate 113K additional demonstrations from Gemini 3 Flash, featuring dense visual grounding and step-by-step
reasoning on the hardest unsolved problems. Experiments on Qwen3-VL show that PRISM consistently improves downstream RLVR performance across multiple RL
algorithms (GRPO, DAPO, GSPO) and diverse
multimodal benchmarks, improving average accuracy by +4.4 and +6.0 points over the SFT-to-RLVR baseline on 4B and 8B, respectively. Our code, data, and model checkpoints are publicly available at https://github.com/XIAO4579/PRISM.
Read the full abstract ▾
All authors Sudong Wang, Weiquan Huang, Xiaomin Yu, Zuhao Yang, Hehai Lin, Keming Wu, Chaojun Xiao, Chen Chen, Wenxuan Wang, Beier Zhu, Yunjian Zhang, Chengwei Qin
Computation and Language · Cross-listed from cs.AI
Mapping the Methodological Space of Classroom Interaction Research: Scale, Duration, and Modality in an Age of AI
Dorottya Demszky, Edith Bouton, Alison Twiner et al. (+2 more)
Research on classroom interaction has long been divided between large-scale observation and in-depth ethnographic work. We propose a framework mapping this methodological space along three dimensions–scale, duration, and modality–where a study's position shapes what it reveals and obscures. We illustrate it through contrasting studies of dialogic teaching–Howe et al. (2019) and Snell and Lefstein (2018)–and an interview with the lead researchers, organized around three questions: what can be operationalized, what mechanisms become visible, and what translates to practice. We then examine how AI is expanding this space and how the framework can guide research and tool design.
Read the full abstract ▾
All authors Dorottya Demszky, Edith Bouton, Alison Twiner, Sara Hennessy, Richard Correnti
Computation and Language
TopBench: A Benchmark for Implicit Predictive Reasoning in Tabular Question Answering
An-Yang Ji, Jun-Peng Jiang, De-Chuan Zhan et al. (+1 more)
Large
Language Models (LLMs) have advanced Table Question Answering, where most queries can be answered by extracting information or simple aggregation. However, a common class of real-world queries is implicitly predictive, requiring the inference of unobserved answers from historical patterns rather than mere
retrieval. These queries introduce two challenges: recognizing latent intent and reliable predictive
reasoning over massive tables. To assess LLMs in such Tabular questiOn answering with implicit Prediction tasks, we introduce TopBench, a
benchmark consisting of 779 samples across four sub-tasks, ranging from single-point prediction to decision making, treatment effect analysis, and complex filtering, requiring models to generate outputs spanning
reasoning text and structured tables. We evaluate diverse models under both text-based and agentic workflows. Experiments reveal that current models often struggle with intent recognition, defaulting to just lookups. Deeper analysis identifies that accurate intent disambiguation serves as the prerequisite for leading these predictive behaviors. Furthermore, elevating the upper bound of prediction precision requires the integration of more sophisticated modeling or
reasoning capabilities.
Read the full abstract ▾
All authors An-Yang Ji, Jun-Peng Jiang, De-Chuan Zhan, Han-Jia Ye
Computation and Language
Repetition over Diversity: High-Signal Data Filtering for Sample-Efficient German Language Modeling
Ansar Aynetdinov, Patrick Haller and Alan Akbik
Recent research has shown that filtering massive English web corpora into high-quality subsets significantly improves training efficiency. However, for high-resource non-English
languages like German, French, or Japanese, aggressive filtering creates a strategic dilemma: should practitioners prioritize diversity by training once on large amounts of lightly filtered web data, or prioritize quality by strictly filtering for a high-quality core and repeating it over multiple epochs? We investigate this trade-off for German by applying hierarchical quality filters to 500M web documents, comparing multi-epoch training on filtered subsets against single-pass training on a diverse corpus. Our experiments across multiple model scales and
token budgets show that repeating high-quality data consistently outperforms single-pass training on larger, less filtered sets. Notably, the performance gap persists even after 7 epochs. Our findings suggest that for non-English LLMs, semantic concentration through quality filtering offers a more viable path to efficient
language modeling than simply maximizing unique data volume. We release our German
language models (called Boldt), as well as our cleaned evaluation
benchmarks to the research community. Our experiments indicate that they achieve state-of-the-art results despite training on 10-360x fewer
tokens than comparable models.
Read the full abstract ▾
Computation and Language · Cross-listed from cs.DL
Measuring research data reuse in scholarly publications using generative artificial intelligence: Open Science Indicator development and preliminary results
Lauren Cadwallader, Iain Hrynaszkiewicz, parth sarin et al. (+1 more)
Numerous metascience studies and other initiatives have begun to monitor the prevalence of open science practices when it is more important to understand the 'downstream' effects or impacts of open science. PLOS and DataSeer have developed a new LLM-based indicator to measure an important effect of open science: the reuse of research data. Our results show a data reuse rate of 43%, which is higher than established bibliometric techniques. We show that data reuse can be measured at scale using LLMs and generative artificial intelligence. The positive effects of research data sharing and reuse may currently be underestimated.
Read the full abstract ▾
All authors Lauren Cadwallader, Iain Hrynaszkiewicz, parth sarin, Tim Vines
Computation and Language
Stable Behavior, Limited Variation: Persona Validity in LLM Agents for Urban Sentiment Perception
Neemias B da Silva, Rodrigo Minetto, Daniel Silver et al. (+1 more)
Large
Language Models (LLMs) are increasingly used as proxies for human perception in urban analysis, yet it remains unclear whether persona prompting produces meaningful and reproducible behavioral diversity. We investigate whether distinct personas influence urban sentiment judgments generated by
multimodal LLMs. Using a factorial set of personas spanning gender, economic status, political orientation, and personality, we instantiate multiple
agents per persona to evaluate urban scene images from the PerceptSent
dataset and assess both within-persona consistency and cross-persona variation. Results show strong convergence among
agents sharing a persona, indicating stable and reproducible behavior. However, cross-persona differentiation is limited: economic status and personality induce statistically detectable but practically modest variation, while gender shows no measurable effect and political orientation only negligible impact.
Agents also exhibit an extremity bias, collapsing intermediate sentiment categories common in human annotations. As a result, performance remains strong on coarse-grained polarity tasks but degrades as sentiment resolution increases, suggesting that simple label-based persona prompting does not capture fine-grained perceptual judgments. To isolate the contribution of persona conditioning, we additionally evaluate the same model without personas. Surprisingly, the no-persona model sometimes matches or exceeds persona-conditioned agreement with human labels across all task variants, suggesting that simple label-based persona prompting may add limited annotation value in this setting.
Read the full abstract ▾
All authors Neemias B da Silva, Rodrigo Minetto, Daniel Silver, Thiago H Silva
Computation and Language
Ease of dependency distance minimization in star-like structures
Emília Garcia-Casademont and Ramon Ferrer-i-Cancho
The syntactic structure of a sentence can be represented as a tree where edges indicate syntactic dependencies between words. When that structure is a
star, it has been demonstrated that the head should be placed in the middle of the linear arrangement according to the principle of syntactic dependency distance minimization. However, hubs of
stars tend to be put at one of the ends, against that principle. Here we address two questions: (1) How difficult is it to minimize dependency distance? (2) Why anti dependency distance minimization effects have been found in
star structures but not in path structures? The ease of optimization is determined by the shape of the optimization landscape. It was demonstrated that the landscape of
star structures is quasiconvex (Ferrer-i-Cancho 2015,
Language Dynamics and Change). As for (1), here we show that it is indeed convex (a particular case of quasiconvexity) both for
star trees and quasistar trees and thus the distance-based optimization problem is simpler than previously believed. As for (2), we argue that (a) competing principles, rather than the difficulty of optimization, must be the actual reason for anti-dependency distance minimization effects and that (b) dependency distance minimization on
star-like structures is less rewarding compared to other structures.
Read the full abstract ▾
Computation and Language
Models Recall What They Violate: Constraint Adherence in Multi-Turn LLM Ideation
Garvin Kruthof
When researchers iteratively refine ideas with large
language models, do the models preserve fidelity to the original objective? We introduce DriftBench, a
benchmark for evaluating constraint adherence in multi-turn LLM-assisted scientific ideation. Across 2,146 scored
benchmark runs spanning seven models from five providers (including two open-weight), four interaction conditions, and 38 research briefs from 24 scientific domains, we find that iterative pressure reliably increases structural complexity and often reduces adherence to original constraints. A restatement probe reveals a dissociation between declarative recall and behavioral adherence, as models accurately restate constraints they simultaneously violate. The knows-but-violates (KBV) rate, measuring constraint non-compliance despite preserved recall, ranges from 8% to 99% across models. Structured checkpointing partially reduces KBV rates but does not close the dissociation, and complexity inflation persists. Human validation against blind raters confirms that the LLM judge under-detects constraint violations, making reported constraint adherence scores conservative. Sensitivity analyses confirm the findings are robust to temperature (0.7 vs.\ 1.0) and pressure type (novelty vs.\ rigor). We release all briefs,
prompts, rubrics, transcripts, and scores as an open
benchmark.
Read the full abstract ▾
Computation and Language
Reliable Answers for Recurring Questions: Boosting Text-to-SQL Accuracy with Template Constrained Decoding
Smit Jivani, Sarvam Maheshwari and Sunita Sarawagi
Large
language models (LLMs) have revolutionized Text-to-SQL generation, allowing users to query structured data using natural
language with growing ease. Yet, real-world deployment remains challenging, especially in complex or unseen schemas, due to inconsistent accuracy and the risk of generating invalid SQL. We introduce Template Constrained Decoding (TeCoD), a system that addresses these limitations by harnessing the recurrence of query patterns in labeled workloads. TeCoD converts historical NL-SQL pairs into reusable templates and introduces a robust template selection module that uses a fine-tuned natural
language inference model to match or reject queries efficiently. Once the template is selected, TeCoD enforces it during SQL generation through grammar-constrained decoding, implemented via a novel partitioned strategy that ensures both syntactic validity and efficiency. Together, these components yield up to 36% higher execution accuracy than in-context learning (ICL) and 2.2x lower latency on matched queries.
Read the full abstract ▾
Computation and Language · Cross-listed from physics.soc-ph
Universal statistical laws governing culinary design
Ganesh Bagler, Gopal Krishna Tewari, Aditya Raj Yadav et al. (+5 more)
Cooking is a cultural expression of human creativity that transcends geography and time through the orchestration of ingredients and techniques, much like
languages do through words and syntax. Yet, beneath the apparent diversity of culinary traditions, whether recipes obey statistical laws comparable to those of other symbolic systems remains unknown. Here we analyze a large corpus of traditional recipes spanning global cuisines, annotated using a state-of-the-art named entity recognition
algorithm into ingredients, cooking techniques, utensils, and other culinary attributes. We find that ingredient usage exhibits Zipf-like rank-frequency scaling, that culinary diversity grows sublinearly with corpus size in accordance with Heaps' law, and that recipe complexity follows Menzerath-Altmann-type relations between the number and average information of constituent units. Consistent with observations in packaged foods, macronutrient concentrations across recipes also display a log-normal signature. Minimal generative models based on preferential reuse, constrained sampling, and incremental modification recapitulate these regularities, suggesting generic processes that shape recipe architecture across cultures. Together, these findings establish recipes as a compositional symbolic system in which complex structure emerges from simple, constrained generative processes.
Read the full abstract ▾
All authors Ganesh Bagler, Gopal Krishna Tewari, Aditya Raj Yadav, Akshat Singh, Pranay Bansal, Ujjval Dargar, Mansi Goel, Madhvi Kumari Sinha
Computation and Language · Cross-listed from cs.LG
Latent-GRPO: Group Relative Policy Optimization for Latent Reasoning
Jingcheng Deng, Zihao Wei, Liang Pang et al. (+4 more)
Latent
reasoning offers a more efficient alternative to explicit
reasoning by compressing intermediate
reasoning into continuous representations and substantially shortening
reasoning chains. However, existing latent
reasoning methods mainly focus on supervised learning, and reinforcement learning in latent space remains highly unstable. We study this problem through the lens of Group Relative Policy Optimization (GRPO), and show that directly adapting GRPO to latent
reasoning is fundamentally non-trivial: latent
reasoning changes both the probability density and the sampling mechanism, causing three coupled bottlenecks: absence of intrinsic latent
manifolds, where unconstrained exploration pushes rollouts off the valid latent
manifold; exploration-optimization misalignment, where trajectory-level rewards can induce incorrect
token-level updates; and latent mixture non-closure, where jointly reinforcing multiple correct latent paths can produce an invalid averaged state. To address them, we propose Latent-GRPO, which combines invalid-sample advantage masking, one-sided noise sampling, and optimal correct-path first-
token selection. Across four low-difficulty
benchmarks (e.g., GSM8K-Aug) and four high-difficulty
benchmarks (e.g., AIME), Latent-GRPO improves over its latent initialization by 7.86 Pass@1 points on low-difficulty tasks and surpasses explicit GRPO by 4.27 points on high-difficulty tasks while using 3–4$\times$ shorter
reasoning chains. It also achieves stronger pass@$k$ performance under Gumbel sampling. These results establish Latent-GRPO as an effective approach for stable and efficient latent
reasoning.
Read the full abstract ▾
All authors Jingcheng Deng, Zihao Wei, Liang Pang, Junhong Wu, Shicheng Xu, Zenghao Duan, Huawei Shen
Computation and Language · Cross-listed from cs.AI
MM-StanceDet: Retrieval-Augmented Multi-modal Multi-agent Stance Detection
Weihai Lu, Zhejun Zhao, Yanshu Li et al. (+1 more)
Multimodal Stance Detection (MSD) is crucial for understanding public discourse, yet effectively fusing text and image, especially with conflicting signals, remains challenging. Existing methods often face difficulties with contextual grounding, cross-modal interpretation ambiguity, and single-pass
reasoning fragility. To address these, we propose
Retrieval-Augmented Multi-modal Multi-
agent Stance Detection (MM-StanceDet), a novel multi-
agent framework integrating
Retrieval Augmentation for contextual grounding, specialized
Multimodal Analysis
agents for nuanced interpretation, a
Reasoning-Enhanced Debate stage for exploring perspectives, and Self-Reflection for robust adjudication. Extensive experiments on five
datasets demonstrate MM-StanceDet significantly outperforms state-of-the-art baselines, validating the efficacy of its multi-
agent architecture and structured
reasoning stages in addressing complex
multimodal stance challenges.
Read the full abstract ▾
All authors Weihai Lu, Zhejun Zhao, Yanshu Li, Huan He
Computation and Language
DPN-LE: Dual Personality Neuron Localization and Editing for Large Language Models
Lifan Zheng, Xue Yang, Jiawei Chen et al. (+6 more)
With the widespread adoption of large
language models (LLMs), understanding their personality representation mechanisms has become critical. As a novel paradigm in Personality Editing, most existing methods employ
neuron-editing to locate and modify LLM
neurons, requiring changes to numerous
neurons and leading to significant performance degradation. This raises a fundamental question: Are all modified
neurons directly related to personality representation? In this work, we investigate and quantify this specificity through assessments of general capability impact and representation-level patterns. We find that: 1) Current methods can change personalities but reduce overall performance. 2)
Neurons are multifunctional, connecting personality traits and general knowledge. 3) Opposing personality traits demonstrate distinctly mutually exclusive representation patterns. Motivated by these findings, we propose DPN-LE (Dual Personality
Neuron Localization and Editing), which identifies personality-specific
neurons by contrasting MLP activations between high-trait and low-trait samples. DPN-LE constructs layer-wise steering vectors and applies dual-criterion filtering based on Cohen's $d$ effect size and activation magnitude to isolate mutually exclusive
neuron subsets. Sparse linear intervention on these
neurons enables precise personality control at inference time. Using only 1,000 contrastive sample pairs per trait, DPN-LE intervenes on $\sim$0.5% of
neurons while achieving competitive personality control and substantially better capability preservation across
reasoning tasks. Experiments on LLaMA-3-8B-Instruct and Qwen2.5-7B-Instruct demonstrate the effectiveness and generalizability of our approach.
Read the full abstract ▾
All authors Lifan Zheng, Xue Yang, Jiawei Chen, Chenyan Wu, Jingyuan Zhang, Fanheng Kong, Xinyi Zeng, Xiang Chen, Yu Tian
Computation and Language
Can AI Be a Good Peer Reviewer? A Survey of Peer Review Process, Evaluation, and the Future
Sihong Wu, Owen Jiang, Yilun Zhao et al. (+5 more)
Peer review is a multi-stage process involving reviews, rebuttals, meta-reviews, final decisions, and subsequent manuscript revisions. Recent advances in large
language models (LLMs) have motivated methods that assist or automate different stages of this pipeline. In this survey, we synthesize techniques for (i) peer review generation, including fine-tuning strategies,
agent-based systems, RL-based methods, and emerging paradigms to enhance generation; (ii) after-review tasks including rebuttals, meta-review and revision aligned to reviews; and (iii) evaluation methods spanning human-centered, reference-based, LLM-based and aspect-oriented. We catalog
datasets, compare modeling choices, and discuss limitations, ethical concerns, and future directions. The survey aims to provide practical guidance for building, evaluating, and integrating LLM systems across the full peer review workflow.
Read the full abstract ▾
All authors Sihong Wu, Owen Jiang, Yilun Zhao, Tiansheng Hu, Yiling Ma, Kaiyan Zhang, Manasi Patwardhan, Arman Cohan
Computation and Language
Beyond Semantics: Measuring Fine-Grained Emotion Preservation in Small Language Model-Based Machine Translation
Dawid Wisniewski and Igor Czudy
Preserving affective nuance remains a challenge in Machine Translation (MT), where semantic equivalence often takes precedence over emotional fidelity. This paper evaluates the performance of three state-of-the-art Small
Language Models (SLMs) – EuroLLM, Aya Expanse, and Gemma – in maintaining fine-grained
emotions during backtranslation. Using the GoEmotions
dataset, which comprises Reddit comments across 28 distinct categories, we assess emotional preservation across five European
languages: German, French, Spanish, Italian, and Polish. Specifically, we investigate (i) the inherent capability of these SLMs to retain emotional sentiment, (ii) the efficacy of
emotion-aware prompting in improving preservation, and (iii) the performance of ModernBERT as a contemporary alternative to BERT for
emotion classification in MT evaluation.
Read the full abstract ▾
Computation and Language
Geometry-Calibrated Conformal Abstention for Language Models
Rui Xu, Yi Chen, Sihong Xie et al. (+1 more)
When
language models lack relevant knowledge for a given query, they frequently generate plausible responses that can be
hallucinations, rather than admitting being agnostic about the answer. Retraining models to reward admitting ignorance can lead to overly conservative behaviors and poor generalization due to scarce evaluation
benchmarks. We propose a post hoc framework, Conformal Abstention (CA), adapted from conformal prediction (CP) to determine whether to abstain from answering a query. CA provides finite-sample guarantees on both the probability of participation (i.e., not abstaining) and the probability that the generated response is correct. Importantly, the abstention decision relies on prediction confidence rather than the non-conformity scores used in CP, which are intractable for open-ended generation. To better align prediction confidence with the model's ignorance, we introduce a calibration strategy using representation geometry within the model to measure knowledge involvement in shaping the response. Experiments demonstrate that we improve selective answering significantly with 75 percent conditional correctness.
Read the full abstract ▾
All authors Rui Xu, Yi Chen, Sihong Xie, Hui Xiong