Peer Revenue Play credits. Real papers.

Papers

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

Computer Science Computation and Language

cs.CL Watch Computation and Language

Tuesday 21 July

Computation and Language

Copy Less, Ground More: Overcoming Repetitive Copying in Long-Context Reasoning via Evidence-Aware Reinforcement Learning

Lizhe Fang, Weizhou Shen, Tianyi Tang et al. (+1 more)

Large language models that generate step-by-step reasoning traces have achieved strong performance on complex tasks, and extending them to long-context settings has emerged as an important frontier. However, we identify a critical failure mode in this regime: repetitive copying, where models extensively copy text from the input into their reasoning traces rather than productively solving the problem. We show that this behavior is pervasive across frontier long-context LLMs and intensifies with context length. By separating each prompt into task-relevant key evidence and irrelevant distractor context, we further show that the root cause is insufficient grounding: models copy from the prompt indiscriminately, and those that fail to focus on key evidence are far more likely to answer incorrectly. Motivated by this diagnosis, we propose GEAR (Grounding Evidence-Aware Reward), a reward shaping method that augments the accuracy signal with a grounding reward for overlap with key evidence and a distractor penalty for overlap with irrelevant context. To enable GEAR on natural-language data, we develop an automated pipeline that constructs evidence-annotated training data from arbitrary documents. We validate GEAR across multiple model scales and benchmarks, showing consistent improvements of up to +4.6 average points over standard RL with accuracy-based rewards, with larger gains at longer contexts, while also reducing repetitive copying and thinking length. Our findings suggest that, even as long-context evaluation shifts from simple retrieval toward complex reasoning, accurate grounding in relevant evidence remains an indispensable capability with substantial room for improvement.
Read the full abstract

All authors Lizhe Fang, Weizhou Shen, Tianyi Tang, Yisen Wang

Computation and Language · Cross-listed from cs.AI

Agents in the Wild: Where Research Meets Deployment

Grace Hui Yang, Pranav N. Venkit, Hooman Sedghamiz et al. (+3 more)

Agentic systems large language model (LLM) based architectures capable of reasoning, planning, acting, and coordinating with tools and other agents are rapidly transitioning from research prototypes to production scale deployments across domains such as software engineering, scientific discovery, and finance. While academic work has emphasized benchmarks and algorithmic innovation, deployment raises new challenges around robustness, safety, and reliability. This tutorial brings together researchers and practitioners to explore advances in reasoning and planning, multi agent coordination, and evaluation, highlighting open challenges arising from deployment experience. Through applied case studies in pharmaceutical discovery and financial systems, we analyze common design patterns that make agentic systems successful, and discuss practical mitigation strategies for failure modes, such as verification pipelines, fallback mechanisms, and human in the loop supervision. Attendees will gain a comprehensive view of the field along with concrete design patterns, evaluation checklists, and templates for safe and reliable deployment across industries.
Read the full abstract

All authors Grace Hui Yang, Pranav N. Venkit, Hooman Sedghamiz, Enrico Santus, Victor Dibia, Ioana Baldini

Computation and Language

Selective State-Space Adaptation and Retrieval for Language Model Reasoning

Atahan Dokme and Larry Heck

Low-rank adaptation introduces a static learned update applied identically to every input. The update provides task-level adaptation but does not explicitly represent token-level or instance-level state variation. A family of adapters is proposed that introduces selective state-space control at two complementary granularities. At the token level, MaLoRA (Mamba-modulated low-rank adaptation) makes the adapter's scaling factor a dynamic input-dependent function with recurrent state across tokens, in contrast to the stateless modulators of prior work. The token-level adapter improves over low-rank adaptation. On the other hand, it differentiates tokens by structural role but not by contextual relevance, which motivates placing evidence selection at the context level. At the context level, MaRA (Mamba Retrieval Adapter) tracks cross-segment reasoning state and selects the segments most relevant to the query. State-space controlled retrieval of approximately three million parameters exceeds an eight-billion-parameter dense retriever on supporting-paragraph recall. Although base models perform poorly on the task without adaptation (14 to 25 F1), MaRA recovers the evidence relevance latent in their representations. Across three frozen backbones and two multi-hop reasoning benchmarks, the end-to-end family improves reasoning accuracy on every cell of the 3-by-2 grid, by +6.4 F1 (+10.0% relative) on average over the LoRA baseline.
Read the full abstract

Computation and Language

Two-Level Meta-Rubrics for Evaluating Open-Ended Generation: GAMUT, a Benchmark for Factual Completeness

Xilun Chen, Zhaleh Feizollahi, Ross Goodwin et al. (+5 more)

Rubric-based evaluation of open-ended generation faces a fundamental tension between expressiveness and reliability. Authoring a faithful rubric requires expressing the structure of the space of good answers: open-ended sets of acceptable options, ordered processes, and the relative importance of facts. Grading with the rubric requires a judge to score consistently, and judges are far more reliable on flat, binary checks than on rich structure. We resolve this tension with a two-level meta-rubric framework. A structured meta-rubric captures the grading criteria at authoring time, and fixed mechanical rules compile it into a flat checklist of binary, machine-gradable checks that an LLM judge scores reliably at evaluation time. We instantiate the framework as Gamut (Grounded Assessment of Multimodal Factuality), a benchmark for factual completeness in long-form generation. Gamut comprises 1,813 questions grounded in real wearable imagery across 10 diverse domains, each paired with an evidence-backed rubric verified by expert human annotators. Evaluating 14 frontier and open-weight models, we find Gamut genuinely challenging (best score 58.7% from Gemini 3.1 Pro), highly discriminative, and robust to the choice of judge.
Read the full abstract

All authors Xilun Chen, Zhaleh Feizollahi, Ross Goodwin, Seungwhan Moon, Scott Yih, Pinar Donmez, Babak Damavandi, Luna Dong

Computation and Language · Cross-listed from cs.LG

CircuitKIT : Circuit Discovery, Evaluation, and Application Toolkit for Mechanistic Interpretability

Pratinav Seth, Hem Gosalia, Aditya Kasliwal et al. (+1 more)

Circuit analysis can support not only model explanation but also downstream interventions such as pruning, editing, steering, and selective fine-tuning. However, conducting such analyses currently requires stitching together separate implementations for discovery, evaluation, and intervention, as well as hand-authoring the contrastive prompts required by many discovery methods. This fragmentation makes methods difficult to compare and limits their application beyond canonical tasks. We introduce CircuitKIT, a source-available library that connects the circuit-analysis workflow through a typed, serializable representation. CircuitKIT provides a suite of discovery algorithms, declarative interfaces for mapping structured data into discovery tasks, complementary circuit diagnostics, and downstream application modules. Together, these components provide common infrastructure for conducting and comparing circuit analyses. The library, examples, notebooks, and documentation are released at https://github.com/Lexsi-Labs/CircuitKIT .
Read the full abstract

All authors Pratinav Seth, Hem Gosalia, Aditya Kasliwal, Vinay Kumar Sankarapu

Computation and Language

Prompt Design at Scale: How Format, Instruction Count, and Context Length Shape Instruction Adherence and Hallucination in Large Language Models

Netanel Eliav

Practitioners make three prompt-design decisions with almost no controlled evidence behind them: how to format instructions and context (markdown, plain text, prose, or tabular), how many simultaneous instructions a system prompt can carry before compliance degrades, and how much context a model can hold before recall and honesty degrade. We report two controlled experiments crossing all three factors on one held, contamination-free synthetic corpus (the "Book of Veyra," 8,780 uniquely-named entities, deterministically regenerable from a fixed seed), evaluated across five models. Experiment 1 (960 calls/model) measures instruction-following decay as rule count N grows from 10 to 160, crossed with four formats and system-prompt vs. user-turn placement. Perfect-response rate collapses to zero by N=80 for every model, format, and placement. Placement produces effects at least as large as format at N=160 in most models, but the direction is model-specific. No model shows a reliable markdown advantage; one 35B model favors plain text instead. Experiment 2 (5,520 calls/model) measures recall accuracy, false-premise sycophancy, and absent-fact fabrication across a 2k-to-512k-token context ladder in the same four formats. Recall stays near ceiling through 64-128k tokens, then degrades sharply and format-dependently: one model's accuracy spread reaches 48 points at 128k tokens. Fabrication never occurs (0/5,760 probes), and sycophancy stays negligible (<=8.3%). What rises sharply near each model's context ceiling is outright refusal to answer (0% to 79-90%), distinct from sycophancy or fabrication. Neither pre-registered format ordering holds, and token overhead (+22% to +37% over plain text) further changes which format is preferable where accuracy spread is genuine. We release the full harness, corpus generator, and raw results (VeyraBench): https://github.com/iNetanel/veyrabench
Read the full abstract

Computation and Language

Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs

Alexander Manev

Although Large Language Models (LLMs) demonstrate remarkable multilingual fluency, their internal knowledge representations remain disproportionately biased toward high-resource languages. This leads to cross-lingual factual inconsistency, where they shift their empirical answer distributions based solely on the prompt language. We investigate whether these biases can be mitigated at inference time, forcing an English-prompted model to answer as if it were queried in target languages (German, Spanish, Bulgarian), and evaluate four intervention strategies: zero-shot contextual steering (persona prompting), internal representation manipulation via Contrastive Activation Addition (CAA), and lightweight weight modification via Direct Preference Optimization (DPO) trained on benchmark-derived factual data as well as conceptual generalization data. To assess alignment, we curate a multilingual factual dataset alongside a novel generalization benchmark comprising culturally rooted queries to determine whether factual interventions transfer to broader target-centric preferences. Experiments on Gemma 3 12B Instruct reveal persona prompting to be the strongest overall intervention, balancing efficacy, safety, and out-of-domain generalization. While CAA yields sharp inconsistency benchmark shifts, it is configuration-sensitive and risks knowledge degradation. DPO-based adapters offer permanent, yet narrower and less transferable gains. These findings suggest that cross-lingual inconsistency is at least partly a selection problem, and that simple contextual interventions may outperform more invasive methods for robust, transferable alignment.
Read the full abstract

Computation and Language

MeetingToM: Evaluating Multimodal LLMs on Theory-of-Mind Reasoning in Multi-Party Meetings

Ziyi Wang, Yuhang Wu, Dongxu Piao et al. (+3 more)

Theory of Mind (ToM), the ability to infer other's beliefs, intentions, and states of knowledge, is central to social interaction, yet remains challenging for current Multimodal Large Language Models (MLLMs), especially in multi-party meetings where cues are distributed across speech and behavior. Existing multimodal ToM benchmarks mainly focus on video-grounded question answering over overt, externally verifiable signals, and provide limited coverage of latent social states and group dynamics. We introduce MeetingToM, a benchmark for complex social behavior reasoning in naturalistic multi-party meetings. MeetingToM targets meeting-specific phenomena such as pseudo-consensus, where apparent agreement masks private dissent under social pressure. The benchmark is hierarchically organized to evaluate ToM at increasing levels of social granularity, including (i) subject-level mental state prediction, (ii) dyadic-level addressee understanding, and (iii) group-level consensus reasoning. We provide a unified evaluation protocol and conduct systematic analyses of representative MLLMs, revealing persistent limitations in integrating non-verbal cues, inferring hidden attitudes, and distinguishing genuine consensus from pseudo-consensus. Our results highlight key challenges and establish MeetingToM as a testbed for advancing meeting-grounded ToM in multimodal models.
Read the full abstract

All authors Ziyi Wang, Yuhang Wu, Dongxu Piao, Xingyu Liu, Tianhui Zhou, Miao Liu

Computation and Language

Selection Shapes the Boundary: A Preregistered Replication of Monotonicity and Label Agreement in Unselected NLI Populations

Haram Choi

Prior work on human label variation (HLV) in natural language inference (NLI) has often relied on re-annotation resources that select items by disagreement level. An earlier study (arXiv:2607.15870) found that hypotheses containing non-upward monotonicity operators showed lower label agreement in ChaosNLI (Cliff's delta = -0.284), which is restricted to items whose majority label carries exactly three of five votes. We preregistered a replication of this boundary in the unselected populations that ChaosNLI was drawn from: the SNLI and MultiNLI development sets, using the same operator tagger and a four-level ordinal agreement outcome. The registered prediction fails. All seven contrasts return a positive Cliff's delta (non-upward items agree slightly more, not less), the only significant confirmatory contrast has the opposite sign to the registration, and every effect is far below our smallest effect size of interest (0.10). Robustness checks support the measurement: simulated tagger misclassification shrinks the effects rather than manufacturing them, and a manual re-tagging audit reaches four-class agreement of 0.875 on a fresh 200-item sample. We conclude that the earlier negative boundary is plausibly a structure conditional on low-agreement selection rather than a population-level property, and that HLV structure claims built on selected re-annotation resources should state their selection conditional explicitly.
Read the full abstract

Computation and Language

The Price of Reasoning: Cost-Quality Tradeoffs in Reinforcement Learning for Neural Machine Translation

Michael Jungo and Aixiu An

Reinforcement learning with verifiable rewards (RLVR) has been established as a viable paradigm for the post-training of Large Language Models (LLMs), including downstream tasks, such as Neural Machine Translation (NMT). With the latest research indicating that RLVR could be the preferred training method for translating legal documents due to the induced reasoning capabilities, it raises the question whether it is really attributed to the reasoning or more generally to the training paradigm. We investigate the importance of including the model's reasoning trace in the generated responses during both training and inference by systematically omitting it from one of the phases. Our experiments show that including the reasoning, specifically during inference, has a positive effect on the overall translation quality. Furthermore, we recognise that the reasoning leads to an increase in output tokens, hence we study the cost-quality tradeoff between the increased computational demands and the improved translation quality.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

AdaFlash: Adaptive Speculative Decoding via On-Policy Distilled Diffusion Drafters

Yu-Yang Qian, Hao-Cong Wu, Chen Chen et al. (+4 more)

Speculative decoding, in which a lightweight draft model first generates a draft sequence that is then verified by the target model, has become a prevalent paradigm for accelerating large language model inference. Recent work such as DFlash further boosts drafting efficiency by leveraging diffusion drafters, whose parallel denoising mechanism enables draft generation in a single forward pass. In this work, we uncover a central pitfall of diffusion drafters: the global dependency, arising from bidirectional attention and KV injection in diffusion drafters, is a double-edged sword. On one hand, it endows the model with parallel generation and global contextual modeling capabilities; on the other hand, this global dependency introduces high variance at both the domain-level and the token-level: acceptance rates fluctuate substantially across different domains, and the acceptance probability varies across draft positions. To tackle this issue, we propose the AdaFlash framework, comprising two components: (i) an on-policy distillation (OPD) algorithm with reverse-KL divergence tailored for diffusion drafters, which delivers stable convergence and continuously adapts the drafter to the deployment distribution; and (ii) an adaptive length head that dynamically adjusts the candidate length on the fly, substantially lowering the verification cost of the target model. Experiments demonstrate that AdaFlash consistently improves the speedup during deployment, with especially significant gains under high-concurrency, achieving up to 66% higher average throughput than previous state-of-the-art methods. Our code is available at https://github.com/ZinYY/AdaFlash.
Read the full abstract

All authors Yu-Yang Qian, Hao-Cong Wu, Chen Chen, Jiacheng Sun, Zhenhua Dong, Peng Zhao, Zhi-Hua Zhou

Computation and Language

Beyond Score Prediction: LLM-Based Essay Scoring and Feedback Generation via Reinforcement Learning with Rubric Rewards

Xuefeng Jin, Jiashuo Zhang, Teng Cao et al. (+1 more)

Large language models (LLMs) have been widely applied to automated essay scoring (AES) and automated feedback generation (AFG). However, existing studies rely primarily on prompt engineering or supervised fine-tuning, while systematic research on reinforcement learning (RL) post-training and automated evaluation of feedback quality remains limited. We propose RLAES, a unified LLM framework that jointly optimizes essay scoring and feedback generation through RL. To make feedback quality measurable, interpretable, and usable for training, we introduce Rubric-based Feedback Evaluation (RFE), an essay-grounded feedback evaluation framework comprising 166 fine-grained binary rubric items and an LLM-as-judge. Building on RFE, we propose Adaptive Gated Feedback Optimization (AGFO), which activates rubric-based feedback rewards on demand during RL, reducing evaluation overhead while improving feedback quality. We also propose Adjacent Contrastive Reasoning (ACR) to improve ordinal score calibration by explicitly contrasting adjacent score levels. Experimental results show that the RFE framework captures essay-feedback consistency, exhibits strong pairwise discriminative power, and closely aligns with expert preferences. On the ASAP benchmark, RLAES-AGFO achieves the best scoring performance among LLM-based methods (QWK = 0.803), while maintaining feedback quality comparable to GPT-5.5 and avoiding the feedback degradation observed under score-only RL. Code and datasets are publicly available at https://github.com/hellomuyi/RLAES.
Read the full abstract

All authors Xuefeng Jin, Jiashuo Zhang, Teng Cao, Bin Yang

Computation and Language

MIRA-Ev:A Benchmark for Granular Evidence Detection and Relational Reasoning in Clinical Exams

Iker De la Iglesia, Johanna Ramirez-Romero, Jose Maria Villa-Gonzalez et al. (+3 more)

Clinical NLP evaluation remains dominated by multiple-choice question answering (MCQA), which scores only final-answer accuracy and cannot detect when a model reaches the correct diagnosis while grounding it in irrelevant, absent, or contradictory evidence. We introduce MIRA-Ev, a clinical argument mining benchmark built on Spanish Médico Interno Residente (MIR) licensing-exam cases, re-annotated by expert clinicians with span-level premises, claims, and directed support/attack relations, and released in parallel Spanish (native), English, and Basque versions, the first clinical argumentation resource in Basque. MIRA-Ev organizes evaluation into a three-tier task hierarchy: evidence sentence retrieval, argumentative component extraction, and relation classification.
Read the full abstract

All authors Iker De la Iglesia, Johanna Ramirez-Romero, Jose Maria Villa-Gonzalez, Irune Urroz García, Ander Barrena, Aitziber Atutxa

Computation and Language

Reasoning Before Translation: Enhancing Legal Machine Translation with Structured Reasoning

Aixiu An, Michael Jungo, Eloi Eynard et al. (+4 more)

Neural machine translation (NMT) in the legal domain is a linguistically and conceptually demanding task, primarily due to the complexity of legal language and the high level of precision it requires. The recent emergence of reasoning-capable language models opens new possibilities for tackling such challenges. They add to a set of other previously proposed techniques to enhance the translation quality, which includes supervised fine-tuning and reinforcement learning. In this work, we perform a comparison between these various approaches. More particularly, we evaluate small language models such as Qwen3.5 4B, Qwen3.5 9B, and Gemma 3 12B enhanced with various re-training paradigms and compare their performances against frontier reasoning models. We focus on the Swiss legal system, which – with its unique multilingual statutes – offers a particularly challenging testbed for reasoning-augmented models. Our results show that the quality of small “base” models can be greatly enhanced, and that reinforcement learning with verifiable rewards can be applied to NMT in the legal domain and surpasses the translation quality of supervised fine-tuning. The performance of enhanced small models is close to the one of state-of-the-art reasoning models yet remains inferior. We also note that re-training paradigms yield diminishing returns as model size increase. The code and models are publicly available at https://github.com/aixiuxiuxiu/Legal-MT-SFT-RL.
Read the full abstract

All authors Aixiu An, Michael Jungo, Eloi Eynard, Mark Drenhaus, Andreas Fischer, Jean Hennebert, Sébastien Rumley

Computation and Language

Automated Extraction of Techno-Economic Data from 76,000 Energy System Studies

Maxime Gorres, Jan Göpfert, Patrick Kuckertz et al. (+5 more)

Energy system models guide societally important decisions, but their credibility rests on quantitative assumptions that are difficult to source and audit. Meta-analyses can improve transparency and modeling practices, but the rapid growth of publications makes manual information extraction increasingly impractical. Consequently, databases are updated infrequently and efforts are often duplicated across research groups. Here, we demonstrate the highly accurate automated extraction of quantitative information from 76,000 energy system studies published since 2010. We compile 3.2 million structured quantitative data points together with 20 million associated metadata entries, spanning a broad spectrum of technologies, methodological approaches and system characteristics. Beyond providing input data for models, the resulting FAIR database make the energy systems literature itself analysable. We show where academic assumptions diverge from empirical observed data, and how research priorities vary at scale across technologies, regions and time. To facilitate broad use within the community, the database is provided through an interactive dashboard, enabling users to filter, analyse and download data according to their specific research needs.
Read the full abstract

All authors Maxime Gorres, Jan Göpfert, Patrick Kuckertz, Noor Titan Putri Hartono, Heidi Heinrichs, Jochen Linßen, Iain Staffel, Jann Michael Weinand

Computation and Language

Translation as Augmentation: Effect of Translated Data on Assessment of Difficulty

Yiheng Wu, Jue Hou and Roman Yangarber

Reliable Text Difficulty Assessment is a prerequisite for valid text simplification workflows and personalized learning applications. However, the development of robust assessment models is severely hindered by a critical bottleneck: the scarcity of expert-annotated corpora containing fine-grained difficulty levels (e.g., CEFR), particularly for lower-resource languages. This paper addresses this data scarcity problem in the context of a low-resource European language. We propose a cross-lingual data augmentation strategy that leverages machine translation to transfer labeled resources from high-resource languages to the target low-resource language. We train BERT-based regression models to predict difficulty scores and investigate whether synthetic, translated data can effectively supplement native training sets. Our experiments demonstrate that augmenting scarce native data with machine-translated corpora significantly improves the accuracy of difficulty estimation, offering a viable solution for languages lacking extensive expert annotations.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Supra Cognitive Modes: A Routed Architecture for Agent Memory

Joshua Tobkin and David Yang

Agent-memory workloads mix direct factual lookup, relation-chain and current-state reasoning, and broad synthesis over long histories. We describe Supra Cognitive Modes (SCM), an architecture that maps explicit or automatically selected per-query modes to retrieval and synthesis payloads over one shared ingest substrate. A frozen semantic classifier and runtime gates dispatch queries among fused lexical and dense lookup, graph or iterative multi-hop handling, and stratified long-form synthesis. The substrate combines multi-granularity embeddings, extracted triples, fact-version metadata, and optional asynchronous enrichments. We characterize the deployed configuration on three benchmarks: Long-term Conversational Memory (LoCoMo; n = 1,986), MemoryAgentBench (MAB; n = 3,671), and LongMemEval (n = 500). The reference run records 84.87% on LoCoMo factoid categories and 68.61% on adversarial abstention, 61.49% on MAB across two repetitions, and 86.00% on LongMemEval. A repository-backed reproduction produces similar aggregate scores and supports task- and mode-conditioned failure analysis. Raw baseline outputs, aligned end-to-end timing for LoCoMo and LongMemEval, and complete token ledgers are unavailable; stored rows also omit some final runtime decisions. The results characterize one implemented routed configuration and its diagnostic failure patterns, while source inspection verifies the per-query control interface and shared-substrate design. Causal routing effects, efficiency gains, and statistical significance remain outside the available evidence.
Read the full abstract

Computation and Language

DAIS: Dependency-Aware Intermediate QA Supervision for Complex Reasoning

Yu Wang, Ming Fan, Xicheng Zhang et al. (+5 more)

Chain-of-thought (CoT) supervision exposes intermediate rationales, but flat rationale targets usually optimize a single reasoning sequence and provide limited supervision on how local conclusions should support later decisions. We introduce Dependency-Aware Intermediate QA Supervision (DAIS), a training-time framework that converts filtered teacher rationales into stage-level QA records. Each intermediate record predicts a local answer conditioned on the previous states needed for that decision, while the final-answer record keeps the original task format; evaluation therefore uses only the original input and optional context. Across GDPR, AIACT, MedQA, and FOLIO with multiple Qwen backbones, DAIS improves average final-answer accuracy over answer-only, flat chain-of-thought, and independent-QA baselines. On policy-compliance benchmarks, it achieves a largest gain of 5.6% and an average gain of 4.2% over the strongest non-DAIS baseline. Controlled ablations show that valid previous-state conditioning contributes beyond longer targets or additional intermediate text, supporting dependency-conditioned intermediate QA as a lightweight auxiliary supervision signal for standard final-answer inference.
Read the full abstract

All authors Yu Wang, Ming Fan, Xicheng Zhang, Zhiyong Li, Zhihu Wang, Caiyue Xu, Dahai Hu, Ting Liu

Computation and Language

Benchmarking Human and Automatic Speech Recognition of Diverse Speech: Initial Results

Ilse Huisman, Rares Popa, Yuanyuan Zhang et al. (+1 more)

Humans are often considered to be the best listeners and seen as the upper-bound performance of automatic speech recognition (ASR) systems. We present a preliminary comparison of the performances of state-of-the-art ASR systems and Dutch native listeners on the recognition of "diverse" speech, specifically Dutch child and older adults' speech and Flemish. Google Telephony outperformed the other ASR systems. Importantly, the ASR systems showed similar performance to the listeners, and in specific cases even outperformed them. Slight performance differences between the listeners and ASR systems were found related to speaker's age and regional accents and utterance length. Future research should focus on making ASR systems more robust to acoustic variability related to aging and regional accents. A comparison of ASR recognition performances on the test stimuli and the full Jasmin-CGN test sets showed the influence of the specific test sets on the conclusions regarding benchmarking human and ASR performance.
Read the full abstract

All authors Ilse Huisman, Rares Popa, Yuanyuan Zhang, Odette Scharenborg

Computation and Language

Content is What Remains: Invariant Speech Tokenization from Parallel Utterances

Laurin Wagner, Bernhard Thallinger, Miroslav Stankovic et al. (+1 more)

Discrete speech tokenizers aim to disentangle semantic from acoustic information, yet targets from self-supervised learning (SSL) models like HuBERT retain non-linguistic variation: speaker identity, prosody, and channel conditions leak into the tokens, inflating entropy. Our key insight is that when enough speakers utter the same words under varying conditions, linguistic content is the only shared factor. We propose PINT (Parallel INvariant Tokenization), which fine-tunes an SSL encoder with alignment losses across parallel utterances and augmentations to distill this shared residual. PINT collapses identical words onto consistent token sequences, drastically reducing conditional entropy. Unlike ASR text, PINT tokens preserve frame-level temporal grounding and serve as drop-in semantic targets for audio codecs. Experiments show a 98.7% relative reduction in speaker probe accuracy (93.1% to 1.2%), a 42% lower ABX error rate, and 27-30% lower LM perplexity versus baselines, confirming that the right invariance is key to efficient learning.
Read the full abstract

All authors Laurin Wagner, Bernhard Thallinger, Miroslav Stankovic, Mario Zusag

Computation and Language

Computational Humor with Multimodal LLMs: Methods, Datasets, Evaluation, and Challenges

Tuo Liang, Zhe Hu, Disheng Liu et al. (+2 more)

Multimodal humor in memes, cartoons, and comics remains difficult for AI systems because intended meaning depends on non-literal mechanisms, shared cultural knowledge, and communicative intent rather than literal scene description. This survey focuses on visual humor understanding in single-image and multi-panel artifacts, while treating humor generation as an emerging downstream frontier. We position the literature against prior humor, sarcasm, and general MLLM surveys and organize it using a capability-centric hierarchy spanning recognition, interpretation and reasoning, and generation. Under this lens, we synthesize benchmark design, evaluation protocols, and modeling paradigms, tracing the field's shift from task-specific fusion models to large-model approaches based on multimodal alignment, evidence-grounded reasoning, and controlled generation. We conclude by highlighting the main barriers to progress: shortcut-prone evaluation, limited cultural and narrative coverage, weak evidence grounding, and unresolved safety and ownership concerns.
Read the full abstract

All authors Tuo Liang, Zhe Hu, Disheng Liu, Jing Li, Yu Yin

Computation and Language

MedDDC-Eval: Diagnosis-Decoupled Evaluation of Multi-Turn Medical Consultation Agents

Guofeng Zhang, Yizeng Quan, Huaiyi Fang et al. (+6 more)

Evaluating multi-turn medical consultation agents requires judging the diagnostic support provided by the histories they elicit through interaction. Yet coupled evaluation lets each policy both elicit the history and generate the terminal diagnosis, so a diagnosis score confounds the elicited history with the policy's own terminal diagnosis generator. We introduce MedDDC-Eval, a diagnosis-decoupled evaluation testbed over held-out cases derived from medical records and online consultations. It applies the same frozen shared diagnostic reader to every policy-elicited history, holding terminal diagnosis generation fixed across policies and enabling comparison under the shared diagnostic reader. It reports diagnostic support, information-acquisition coverage, and efficiency. LLM-assisted semantic matching followed by deterministic one-to-one assignment makes the diagnosis-trajectory-efficiency (D/T/E) scores auditable. In a fixed-history audit across eight policies, replacing each policy's own generator with the shared diagnostic reader shifts diagnosis F1 by 2.2-19.0 points and reverses 18% and 36% of pairwise orderings on the Record and Dialogue splits. To examine downstream utility, we use standard Group Relative Policy Optimization (GRPO) with a separate training-time reward that targets the same diagnosis and trajectory dimensions. Relative to its Qwen3-32B initialization, the trained policy gains 9.6 and 4.6 aggregate-score points on the held-out Record and Dialogue splits, respectively, and ablating either feedback signal reduces the aggregate score on both. Together, MedDDC-Eval supports comparison under a shared diagnostic reader and evaluation-informed policy development, while complementing end-to-end evaluation when terminal diagnosis generation is also part of the target capability.
Read the full abstract

All authors Guofeng Zhang, Yizeng Quan, Huaiyi Fang, Jianwei Lv, Jinyao Liu, Xunxu Duan, Lening An, Yu Ouyang, Junfeng Wang

Computation and Language

Disentangling Curriculum Learning in NLP: Towards a Unifying Taxonomy

Vanessa Toborek, Florian Seiffarth, Sebastian Müller et al. (+1 more)

Despite more than a decade of curriculum learning (CL) research in NLP, the field lacks a principled account of which difficulty function or scheduler to use for a given problem. To understand what has hindered progress towards this account, we propose a fine-grained taxonomy separating difficulty evaluation from training scheduling to enable systematic analysis of CL strategies. For difficulty evaluation, we distinguish attribution source and task dependence, revealing difficulty as a perspectival concept encoding different assumptions about what makes an instance hard to learn. For scheduling, we provide the first formalisation of CL schedulers in terms of expected training contribution, enabling comparison across implementations by introducing retention regimes and monotonicity properties. Applied in a dedicated analysis of CL works in NLP, our taxonomy reveals a systematic incomparability problem: prior works conflate distinct notions of difficulty and scheduling, often pursuing different objectives under the same CL label – hindering comparison and the accumulation of a coherent evidence base. Beyond diagnosis, the taxonomy supports the design, analysis, and comparison of CL strategies, and motivates evaluation practices that disentangle the sources of observed improvement.
Read the full abstract

All authors Vanessa Toborek, Florian Seiffarth, Sebastian Müller, Tamás Horváth

Computation and Language

AutoJourn: Multi-Perspective Summarisation, Bias Detection and Bias Neutralisation for LLM-Generated News in Automated Journalism

Himel Ghosh, Ahmed Mosharafa and Georg Groh

We present AutoJourn, a demonstration system for multi-perspective news generation and bias-aware evaluation using large language models (LLMs). The system tackles three core challenges in responsible automated journalism: extracting diverse perspectives from unstructured social media discussions, generating summaries that preserve viewpoint diversity, and detecting or mitigating bias in AI-generated news. The pipeline integrates advanced prompt engineering with optional retrieval augmentation to produce semantically diverse perspective sets, a multi-perspective summarisation module that merges conflicting viewpoints into balanced summaries, and a bias analysis suite supporting sentence-level bias detection and type classification in the generated news article, and automatic neutralisation. Users can inspect perspective clusters, compare stance-specific summaries, generate news articles, and apply bias-aware rewrites directly in the interface. We evaluate each component with intrinsic metrics – semantic diversity, summary quality, and bias reduction and show improvements over strong baselines while maintaining content fidelity. A live, publicly accessible demo accompanies the paper to facilitate reproducibility and further research on socially responsible automated journalism.
Read the full abstract

Computation and Language

Verifiable Self-Evolution for Open-Ended Dialogue Skills via Future-Feedback Prediction

ChaoJin Zhao and Xuan Jiang

Textual skills provide a lightweight way to improve frozen language-model agents, but their self-evolution normally requires a stable validation signal. Such signals are natural in mathematics or code, where an answer can be checked after it changes, yet are problematic in open-ended dialogue: changing the assistant response also changes the user's next reaction, so a logged reaction cannot directly evaluate a counterfactual response. We propose future-feedback skill evolution, which first redirects self-evolution from prescribing the current answer to predicting whether the observed answer will lead to a positive or negative subsequent user signal. This prediction task is verifiable on fixed logged tuples and therefore supports validation-gated textual optimization. The evolved feedback skill captures interpretable criteria for response quality and can subsequently serve as a diagnostic and optimization target for answer skills. On a proprietary, privacy-preserving sales-assistant dataset, careful quality filtering and a balanced resolved/unresolved split yield more than 75% prediction accuracy. Beyond this result, the central contribution is a formulation that converts otherwise moving conversational feedback into a fixed offline learning target, enabling reproducible skill evolution without placing every candidate skill in live traffic. We discuss the boundary between observational verification and counterfactual validity, and position the method as an offline optimization stage rather than a replacement for final human or online evaluation.
Read the full abstract

Computation and Language · Cross-listed from cs.AI

Measuring Reward-Seeking via Contrastive Belief Updates

Axel Højmark, Jérémy Scheurer, Evgenia Nitishinskaya et al. (+5 more)

Language models trained with reinforcement learning may learn to optimize the grader's judgment rather than the intended objective. This "reward-seeking" is difficult to measure because a model that pursues the grader's judgment and one that pursues the intended objective behave identically whenever the grader rewards the intended behavior. We measure reward-seeking using Contrastive Synthetic Document Finetuning to change a model's beliefs about what the grader rewards, putting those beliefs in conflict with what users or developers want, and measuring the rate at which the model adopts each party's preferred behavior. Applied to intermediate checkpoints of a capabilities-focused OpenAI o3 RL run, without safety training, we find that these checkpoints often side with grader preferences over those of users or developers on coding and alignment tasks. This tendency to side with the grader trends upward throughout RL training. For example, in an environment that forces a choice between keeping a promise to a supervisor and breaking it to complete the task, a late capabilities-focused o3 checkpoint breaks the promise 87% of the time when SDF documents say the grader rewards task completion, versus 9% when they say it rewards honesty (a choice its chain-of-thought often makes explicit). An earlier checkpoint is far less sensitive (40% vs. 24%). Our method also generalizes to reward-hacking models. A model organism trained to reward-hack (gpt-oss-120b) is more than twice as sensitive to grader preferences as the unmodified model, with the mean behavioral shift in favor of the grader rising from 33% to 86%. These results indicate that RL can increase reward-seeking over the course of training, producing models that may act against their developers' intentions when they believe that doing so leads to higher reward.
Read the full abstract

All authors Axel Højmark, Jérémy Scheurer, Evgenia Nitishinskaya, Felix Hofstätter, Jason Wolfe, Theodore Ehrenborg, Bronson Schoen, Alexander Meinke

Computation and Language · Cross-listed from cs.LG

H$^2$SD: Hybrid Hindsight Self-Distillation

Qiye Cai, Yichuan Ma, Weitao Ma et al. (+6 more)

Reinforcement learning with verifiable rewards (RLVR) provides reliable outcome supervision for language model reasoning, but scalar trajectory rewards offer limited guidance for token-level credit assignment. Existing self-distillation methods incorporate privileged teacher information through fixed strategies: direct distribution matching may destabilize successful behavior, while magnitude modulation provides limited corrective guidance after failure. We propose Hybrid Hindsight Self-Distillation (H$^2$SD), which adapts both teacher context and update strategy to trajectory correctness. For successful trajectories, the teacher conditions on the verified student response and a rephrasing instruction to rescore the original tokens. These scores modulate reward-derived updates while preserving advantage signs. For failed trajectories, a verified reference hint guides corrective supervision through reverse KL distillation. Across Sudoku, Calcudoku, Arrow Maze, and mathematical reasoning with two Qwen3 models, H$^2$SD achieves an overall average accuracy of 43.33%, outperforming the strongest baseline by 11.43 percentage points while maintaining stable optimization.
Read the full abstract

All authors Qiye Cai, Yichuan Ma, Weitao Ma, Peiji Li, Yongkang Chen, Yicheng Zou, Linyang Li, Xiaocheng Feng, Bing Qin

Computation and Language

Constrained CTC Decoding for Efficient Diacritic Restoration

Rufael Marew, Amr Keleg and Hanan Aldarmaki

In this work, we address diacritic restoration for Arabic speech transcripts. Most speech data are undiacritized, limiting the ability of modeling fine-grained phonological distinctions. The speech modality has recently been explored as a way to complement text-based diacritic restoration efforts. We propose an efficient non-autoregressive approach for speech-to-text diacritization based on Connectionist Temporal Classification (CTC). Our method incorporates hard constraints during decoding by constructing a character-level diacritization lattice from an undiacritized transcript and restricting hypotheses to valid diacritized realizations. We evaluate on Classical Arabic and Modern Standard Arabic test sets (namely, ArVoice and ClArTTS) against a more computationally-complex multi-modal diacritic restoration baseline, and show statistically significant reductions in diacritic error rates in both, demonstrating that the proposed approach offers both performance and efficiency gains.
Read the full abstract

Computation and Language

Transcription Policy as a Latent Variable: Activating Controllable Verbatim ASR with Word-Level Timing

Laurin Wagner, Mario Zusag and Bernhard Thallinger

Modern ASR models trained on heterogeneously annotated data treat transcription style (verbatim vs. intended) as an uncontrolled latent variable, causing measurable decoding instability, evaluation confounding (up to 60% of reported WER attributable to style mismatch), and unreliable word-level timing. We show that models already encode both styles; the challenge is controlled activation. Using coverage-aware decoder task tokens trained on parallel verbatim/intended transcript pairs, we raise German disfluency F1 from 10% to 79% zero-shot, despite English-only training. Full English-only fine-tuning surpasses all baselines in verbatim accuracy, disfluency detection, and intended-mode quality across both languages. We further introduce supervised cross-attention fine-tuning that improves word-level timestamps on disfluent speech beyond forced-alignment baselines. Finally, we propose verbatimize, a new task enabling scalable creation and enrichment of speech corpora with high-quality canonical verbatim transcriptions.
Read the full abstract

Computation and Language

Reasoning Error from Known Fact: Step-Level Self-Consistency Group Relative Policy Optimization for LLMs

Xiaomeng Hu, Jiaqi Hu, Hao Chen et al. (+4 more)

With the rapid advancement of large language models (LLMs), modern systems not only possess strong foundational capabilities and extensive knowledge, but can also solve complex problems via long, multi-step reasoning. However, as reasoning traces become longer, LLMs may produce a substantial amount of hallucinated content during the reasoning process, which is often difficult to detect. In this work, we conduct a fine-grained analysis of hallucinations arising in LLM reasoning and find that the reasoning traces are particularly prone to Context-Sensitive Factual Hallucinations: cases where the model actually has the relevant knowledge, yet makes factual errors due to contextual interference during reasoning. To address this issue, we propose Step-level Self-Consistency Group Relative Policy Optimization (SSC-GRPO), which assigns step-level rewards to reasoning traces by computing self-consistency scores of individual steps across multiple rollouts. Compared with prior methods, SSC-GRPO achieves state-of-the-art performance on both mathematical reasoning benchmarks and hallucination leaderboards. Our results offer a new perspective for detecting and mitigating hallucinations in the reasoning process of large language models.
Read the full abstract

All authors Xiaomeng Hu, Jiaqi Hu, Hao Chen, Qi Zhang, Zhanming Shen, Wentao Ye, Junbo Zhao