Papers

Memes play a central role in modern online communication, especially in contexts of political conflict where visual and textual cues combine to convey persuasive messages. We study persuasive technique detection in Ukrainian memes using vision-language models (VLMs). We introduce UkrMeme, a dataset of 200 organically produced Ukrainian wartime memes annotated for persuasion techniques, and PTM-UA, a Ukrainian translation of a subset of the Propaganda 950 benchmark that preserves original visual content while translating meme text. We evaluate VLMs in a zero-shot prompting setup across image, text, and multimodal inputs using hierarchical F1 metrics. Finally, we compare Layer-wise Relevance Propagation (LRP) and model-generated token attributions and analyze the semantic categories of evidence emphasized by each explanation method.

On-policy self-distillation promises to replace the sparse outcome signal of verifiable-reward RL with dense token-level credit: a teacher conditioned on the reference solution scores the student's own rollout, and the resulting distributional shift is distilled back into the policy. We investigate whether this scoring signal actually contains corrective credit. Using gold error spans in student rollouts, we test the two signals consumed by self-distillation objectives, the full-distribution KL and the token log-ratio, for two properties: whether they localize errors and whether they push probability away from erroneous student tokens. They exhibit neither of these properties nor can they reliably distinguish correct from incorrect rollouts within the same problem. We argue that this failure is structural: locating an error in a reasoning trace is itself a search problem, while a scoring teacher is never explicitly asked to perform that search. The same model, with the same privileged information, localizes errors substantially better when allowed to generate a critique. This forms the motivation for critique-conditioned self-distillation: first generate a critique of the current rollout, then condition the scoring teacher on that critique. The resulting signal localizes errors, is corrective within error spans, and shifts probability toward tokens that initiate backtracking such as “Wait”, “But”, and “Actually”. We further show that the critique is useful both at test time, where resuming a wrong rollout with the critique in context converts math problems the model cannot solve by sampling, and in cold-start GRPO, where it provides an advantage signal on prompts for which verifiable rewards alone provide none.

Legal language understanding requires reasoning over semi-structured regulatory documents, yet standard pre-training objectives do not explicitly target this structural capability. Because current legal language models rely solely on masked language modeling, they treat statutory text as flat token sequences and ignore the underlying hierarchical constraints. To bridge this gap, we present a multi-task pre-training approach for German traffic law (StVO). Our approach combines three complementary objectives: Next Sentence Prediction (NSP), Knowledge Graph (KG) Guided Clustering (CLU), and Knowledge Graph (KG) Aware Paragraph Classification (CLS). This approach explicitly encodes structured legal knowledge directly into the model weights. This encoding enables zero-shot legal reasoning at inference time and removes the need for discrete graph traversal. We evaluate our approach across nine Transformer-based encoders and 324 experimental runs. On synthetic StVO question answering, our models demonstrate consistent relative improvements between 10.3% and 155.0% over baselines. Furthermore, relative to the established LexGLUE baselines (Chalkidis et al., 2022), our models improve over prior results on four of seven legal benchmarks. It also achieves consistent zero-shot transfer gains between 3.4% and 15.8% on Austrian driving regulations. These experimental results indicate that our training approach generalizes successfully beyond the initial training legal domain.

Long-context serving runs out of KV memory before the hardware reaches its useful throughput. HiSparse stores history in host memory and fetches only the selected entries, but it uses one cache budget at every load and does not handle draft-model speculative decoding such as DSpark and DFlash2. We present Adaptive DHiSparse, which supports speculative decoding over hierarchical sparse KV and adapts residency to load. It protects the required working set, retains additional history while GPU memory is spare, and recovers the rest through exact fused retrieval. DSpark and DFlash2 keep their proposal and acceptance rules: queries share missing-row retrieval, draft KV uses a window-bounded ring, and verification staging is reused across layer groups. At 100K context and concurrency 128, GLM-5.3 with DFlash2 delivers 793 output tokens/s per decode GPU, 1.45$\times$ HiSparse and 1.27$\times$ non-speculative adaptive serving. Held to the same latency bounds as HiSparse, it sustains up to 4$\times$ the concurrency and 2.3 to 3.7$\times$ the output, depending on the per-token bound. On DeepSeek-V4.1-Flash at 500K context and concurrency 512, Adaptive DHiSparse with DSpark delivers 1.85$\times$ the output of resident serving with DSpark. In a load step, optional residency reduces host fetches from 14.2% to 1.4% of selected rows at low load and releases memory under pressure.

Any automation within reliability-critical contexts such as medicine or law has to be transparent and verifiable. Despite their advantages, current LLM-based systems do not meet these requirements due to their inherent black-box nature. Retrieval Augmented Generation (RAG) attempts to ground the generative process in the relevant interpretable documents, thereby reducing the risk of AI 'hallucinations' and potentially enabling verification of the initial generation context. However, as the number and size of potential sources increases, naive retrieval approaches tend to perform worse, ignoring critical context elements. In this paper we aim to tackle this problem in the context of German Legal Code. Using structure-driven parsing to create a Knowledge Graph (KG) for the source document navigation and retrieval, we compare several KG-informed chunking and retrieval strategies and evaluate their effect on the accuracy of retrieved citations.

Prior work suggests fraudulent scientists write their articles with higher rates of linguistic obfuscation than genuine scientists, with a possible mechanism suggesting obfuscation makes research more costly for a reader to analyze. This result has not undergone additional empirical scrutiny in the open science era, nor have other psychological mechanisms been evaluated. The current replication and extension addressed this opportunity by comparing 800 fraudulent and 800 genuine control papers (2013-2025). The results supported the linguistic obfuscation hypothesis, suggesting fraudulent papers were more obfuscated than genuine ones. A parallel mediation model evaluated three competing mechanisms, finding that reality monitoring, as measured through decreased rates of adjectives, best explained the link between fraudulent science and linguistic obfuscation. Cross-validated detection accuracy using obfuscation features (60.7%) and an elastic net of over 100 language dimensions (71.6%) were significantly greater than chance. Implications for scientific fraud and psychology of language research are discussed.

Large language models struggle to reliably identify and correct errors in their own reasoning. Unlike prior training-based approaches that fine-tune an LLM to generate revisions or feedback, we learn where to intervene in failed rollouts and determine how to redirect them. To choose where to intervene, we train a 130K-parameter MLP head on a frozen encoder using only final-answer correctness. For redirection, we find that under a fixed generation budget, the tested inline injections fail to reliably redirect reasoning, whereas presenting the same instruction in a new user turn changes the model's course. The resulting correction framework, UPTURN (User-Prompt TURN injection at a learned position), requires no LLM fine-tuning: it retains a selected prefix, injects an instruction as a new user turn, and regenerates the suffix. Across four reasoners and six math benchmarks, UPTURN improves average accuracy in every setting over a 32-rollout baseline. It outperforms self-correction and reflection-token injection baselines under matched correction budgets, transfers across reasoners and to GPQA-Diamond without retraining, and complements GRPO and self-correction methods.

Many machine learning approaches are two-level: an expert selects a class of methods, within this class the choice of key components is critical and usually manual. This paper addresses the automatic tuning of these components and proposes a framework transferable across models of different nature. The multi-agent large language model (LLM) framework has three roles—analyst, coder, and recorder~—with an experiment log and a sandboxed execution environment. Unlike systems that optimize the entire ML script or search architectures in a discrete space, the framework operates over an open but semantically narrow space: the agent generates code only for the specified components, leaving the overall network structure unchanged. The approach is validated in two directions: tuning a classification spiking neural network with correlation learning, and solving high-order differential equations with Physics-Informed Neural Networks (PINN). Across five tasks—spiking classification on MNIST and Digits, and boundary-value problems for the generalized nonlinear Schrödinger equation (GNLSE) and Kuramoto–Sivashinsky (KS) equations—the agent outperforms the baselines: classification accuracy is 13 percentage points (MNIST) and 6 percentage points (Digits) higher; the error against the analytical single-soliton solution is reduced approximately 2491-fold; the invariant-preservation error drops approximately 325-fold in the selected GNLSE run with a Gaussian perturbation and by more than 100-fold on average over repeated initializations; and the mass drift drops 14-fold for the KS problem with a Gaussian perturbation. The framework thus tunes key components on tasks of different nature without changing its internal structure.

Attention Residuals (AttnRes) enables content-dependent retrieval over network depth, which can be viewed as single-stream softmax attention along the depth axis. Inspired by the success of multi-head attention along the sequence axis, we introduce Multi-Stream Attention Residuals (MSAR), an extension of AttnRes that generalizes this mechanism to multi-head depth attention, yielding substantial performance gains with minimal additional overhead. MSAR draws stream diversity from sequence context through multi-scale causal convolutions, allowing different heads to retrieve historical representations through distinct contextual views while preserving the integrity of the hidden representation. The retrieved streams are combined through lightweight adaptive mixing to generate the single-stream layer input. By retaining single-stream sources and expanding them on demand, MSAR supports multi-stream computation without requiring persistent multi-stream storage or communication. Across all 12 downstream benchmarks at the 18B and 28B MoE scales, MSAR consistently outperforms prior residual variants, including standard residual connections, manifold-constrained Hyper-Connections (mHC), and AttnRes. Scaling-law experiments yield an iso-loss compute speedup of $1.58\times$ over the standard residual baseline, compared with $1.35\times$ for AttnRes and $1.30\times$ for mHC. These results make MSAR a promising residual architecture for large-scale pre-training.

Evaluating large language models (LLMs) on AMLworld's smallest test splits is costly because each of over a million targets requires a separate graph prompt. Sampling reduces this cost, but small balanced subsets change unweighted precision and F1. We introduce AMLworld-Compact, an evaluation set that retains every illicit test transaction and samples benign transactions from supervised-score strata. Its HT-Coreset protocol reduces evaluation targets by 271x and 611x on HI-Small and LI-Small. Horvitz-Thompson weights give unbiased full-test confusion-count estimates for fixed predictions; recall is exact, while precision and F1 are ratio estimates. In a held-out study using Graph Feature Preprocessor (GFP) features, one boosted-tree family defines strata for evaluating the other over 500 draws per configuration. At the released sizes, F1 RMSE is 0.00-2.58 percentage points versus 4.44-7.41 for uniform sampling with HT weights; conservative intervals remain broad. Under the evaluated graph prompts, seven open-weight LLMs have weighted precision near the predict-all-illicit baseline. In an outcome-stratified audit of 1,000 HI-Small few-shot outputs from one run per model, lexical checks pass for 84.5% of correct and 56.2% of incorrect binary verdicts, so these markers do not establish correctness. We release the evaluation sets, construction code, and audit tools at https://huggingface.co/datasets/typhoon-ai/AMLworldCompactEval and https://github.com/nat-nischw/AMLworld-Compact.

Professional agents increasingly work through command-line interfaces and multiple tools over shared workspaces containing files, messages, and business applications. These environments create safety risks that are difficult to capture with attacks placed at a single, known read location. We introduce APEX-Harm, a benchmark for evaluating prompt and script injection against long-horizon agents performing finance, consulting, and legal tasks in large workspaces. Each task contains 115 to 328 files, together with email, calendar, chat, document, spreadsheet, PDF, and code-execution tools exposed through MCP servers. The benchmark includes six attack categories covering static and dynamic injections, task-specific script replacement, and attacker-provided generic scripts, and evaluates attack success, exposure, and task performance. We evaluate 12 frontier models on 180 task instances per model. Opus 5 and Sonnet 5 have the lowest overall attack success rates (8% and 20%), while task-specific script replacement and attacker-provided generic script attacks remain substantially more effective, averaging 81-84% and 73% attack success, respectively. Gemini 3.8 Flash achieves the highest mean task score (0.877) while still exhibiting a 33% attack success rate. These results show that completing the requested task is not sufficient evidence of safe behavior: an agent can produce a correct deliverable while also performing an unauthorized action in the workspace.

User simulators are gaining attention as environments to train LLM assistants that can better cope with the nuances of multi-turn conversations with real users. Prior work on user simulation has emphasized generating realistic user utterances, leaving unexplored their potential as a source of direct feedback on the assistant's responses. Such evaluation signals may not be explicitly communicated in language, but can instead reside in the latent reasoning of users. In this work, we introduce user simulators with intrinsic feedback, namely, models equipped with latent evaluation states: internal thoughts, reactions, and scalar ratings that evaluate the assistant from a first-person user perspective at each conversation turn. We train LLM-based simulators using interaction logs between real users and assistants and demonstrate that intrinsic feedback improves simulation realism and critiquing behaviors versus prompted simulator baselines, while also providing turn-level evaluations that existing fine-tuned simulators lack. Further, evaluations generated by our trained simulators correlate better with human judgments than those of prompted simulators of the same size. Lastly, our experiments reveal that simulators with intrinsic feedback provide effective reward signals for multi-turn RL training, matching setups that rely on external judge LLMs for rewards while cutting RL compute by more than half.

Negative polarity items (NPIs), such as "anymore" and "ever", are grammatical only in select environments, predominantly in the context of negation. Because their grammaticality depends on both syntax and semantics, they have a rich, multilingual literature in theoretical linguistics, yet work on NPIs in language models is almost entirely confined to English. We introduce BLiMPPI, a multilingual evaluation suite of over 12,000 NPI minimal pairs in 33 languages, built by a pipeline that discovers NPIs in Universal Dependencies (UD) treebanks, labels their licensing environments, and perturbs these into minimal pairs. Evaluating 12 multilingual and 2 monolingual models alongside n-gram baselines, we find that multilingual models struggle most with lower-resource languages and with pairs that are not n-gram solvable. We further examine the effects of NPI typology, language family, licensing variability, and NPI-licenser distance. Data: https://github.com/arzuburcuguven/blimppi-data

Dynamic vocabularies built on-the-fly via Lempel-Ziv-Welch (LZW) compression let a language model represent the same text in fewer tokens, but prior work shows that this efficiency comes at a cost to reasoning and perplexity. We show that this quality gap is not a fundamental price of compression but can be substantially closed through targeted improvements to the compression layer and LZW codec, without modifying the underlying pretrained language model. Our analysis reveals four key issues: (1) LZW's compression algorithm creates an implicit conflict between the input embedding and the output embedding, a tension that a single shared embedding cannot resolve; (2) naively applying RoPE to LZW-compressed sequences ignores that merged tokens can span multiple tokens; (3) merging certain tokens (notably digits) disproportionately hurts downstream accuracy; and (4) dynamic vocabularies introduce an inherent segmentation ambiguity that confounds standard evaluation. We address these through decoupled input/output dynamic embedding layer, size-aware RoPE positioning, and a modified LZW codec supporting selective merges and multi-view segmentations. Across four models, our recipe improves perplexity on every corpus we evaluated as well as mathematical reasoning accuracy, closing roughly half the quality gap between zip2zip and the uncompressed baseline at a similar compression ratio on general text. This represents a significant step toward lossless adaptive-vocabulary compression for language models.

Detectors of machine-generated text score almost perfectly on scientific-literature benchmarks, implying we can reliably distinguish with these detectors human and machine-generated documents. We looked at ten scientific-document subsets from six popular resources, and on several of them that apparent ease is a side effect of how the documents were put together. A logistic regression using 27 superficial features reached 0.833-0.997 macro-F1, and on the three arXiv subsets it outperformed every detector we evaluate with F1-scores 0.951-0.997. The strongest cue to the detector is the PDF extraction artifact: human texts retain their original line breaks, whereas generated texts are extracted as a single block of prose. This cue, together with three other formatting features, tell the classes apart at 0.996 AUC. To distinguish benchmarks that contain such a signal from detectors that rely on one, we introduce format counterfactuals: just re-wrapping a machine-written abstract without changing a single word reverses 20-27% of a public detector's correct decisions. Normalizing formatting before training reduces the generalization gap from 0.32 to 0.22. On format-clean benchmarks, it also raises out-of-domain macro-F1 from 0.33 to 0.65 for detectors trained on artifact-containing subsets.

Repository-level issue-resolution benchmarks such as SWE-bench evaluate submitted patches primarily based on whether they pass the correctness tests, leaving the runtime performance of different test-passing patches largely unexplored. Evaluating patch performance is essential but challenging since designing performance workloads for each benchmark instance is costly. However, these benchmarks already provide a natural candidate: the built-in FAIL\_TO\_PASS (F2P) tests, which exercise the behavior affected by the issue fix. We therefore investigate whether the provided F2P tests can be repurposed to evaluate patch performance. Overall, we choose four frontier models, Claude Opus~5, GPT~5.6, DeepSeek V4 Pro, and Qwen~3.8 Max, paired with SWE-agent to resolve the pre-filtered measurable corpus from verified splits of SWE-bench and SWE-bench-Live, yielding 262 cases with at least two different correct solutions, including the developer ground-truth patch. We then measure each solution's runtime performance in instruction count and wall-clock time on the native F2P test, and use Welch ANOVA instead of the RSD proposed in prior work as a more robust statistical discrimination criterion. We find that this methodology exhibits notable discrimination capability in runtime performance and identifies performance differences among patches in almost half of the instances. Our empirical study shows that the runtime performance of generated patches is independent of the issue resolution rate of LLMs, with nearly reversed model rankings. The case study also shows that the performance analysis reveals architectural regression in some patches that the correctness evaluation overlooked. These findings suggest that resolution rate does not adequately characterize LLMs' capability for repository-level issue resolution, motivating future evaluations to consider patch performance alongside it.

Recent speech LLMs can leverage rich paralinguistic cues, enabling more complex and personalized interactions. However, little work investigates how this sensitivity can lead to stereotyping, and existing evaluations largely measure demographic disparities without distinguishing beneficial personalization from stereotypical behavior. We introduce BASIS-Bench, a benchmark designed to disentangle these two aspects through three axes: (1) demographic robustness, measuring disparities in content-based tasks across speakers and demographic classification capabilities; (2) stereotypical bias, measuring whether perceived demographic cues trigger unprompted stereotypical behavior; and (3) personalization, measuring whether models can tailor responses to demographic characteristics when explicitly requested. We evaluate eight recent speech LLMs and find that, despite often recognizing demographic cues, most fail to translate them into appropriate personalization. Explicit demographic labels improve personalization but also amplify stereotypical associations already present in the textual backbones. These findings highlight the need to train speech LLMs not only to perceive demographic cues, but to learn when they should shape their responses.

Multimodal language models (MLLMs) exhibit promising spatial reasoning capabilities, yet the internal mechanisms supporting them remain poorly understood. We uncover S-Space, a linear representation subspace in which object-token representations encode continuous coordinates along horizontal, vertical, and distance axes. Although we identify these axes by differentiating spatial-answer scores with respect to object-token activations, S-Space is intrinsic rather than induced by specific data or prompts, remaining readable under non-spatial questions. We show that S-Space functions as an internal spatial workspace of MLLMs. Perception encodes object locations into it, forming spatial representations that can be directly read out to support accurate spatial judgments. Internal reasoning transforms it for changing viewpoints. Verbal responses selectively access it, evidenced by causal interventions that alter spatial judgments while largely preserving non-spatial responses. Case studies further show that S-Space assigns locations to unseen entities and the observer, and connects physical directions with abstract linguistic meanings, suggesting both the capabilities and limitations of large-scale language training. Finally, we use S-Space as a diagnostic tool for spatial reasoning failures. Direction-word readouts reveal systematic coupling between cardinal and image-relative reference frames. More importantly, a perception-computation decomposition exposes a key bottleneck: MLLMs can form capable spatial representations, yet fall short in the spatial transformations required to reason reliably across viewpoint changes.

Clinical prognosis requires integrating structured measurements, physiological waveforms, longitudinal history, and population context. We introduce PRISM (Population-Retrieved Inference for Structured Multimodal Prognosis), a multimodal language model that combines per-feature numerical tokenization, recency-weighted aggregation of longitudinal ECGs, and a dual-modality population memory bank. Retrieval operates in representation space: the model receives mean-pooled embeddings of similar training-cohort encounters, while no individual retrieved record enters the prompt. On the 1,443-task MDS-ED benchmark, PRISM achieves AUROCs of 86.2% for diagnosis, 93.1% for deterioration, 91.8% for ICU admission, and 92.8% for mortality. These results use a 96-hour pre-arrival window unavailable to reported baselines. With benchmark-window inputs only, PRISM reaches 0.8895 AUROC, compared with 0.8937 for UniPACT and 0.8889 for MDS-ED. A matched-input compact fusion model using the same five evidence streams and frozen encoders reaches 0.7652 AUROC, versus 0.9092 for PRISM. Replacing numerical EHR encoding with text serialization reduces AUROC by 3.66 points; removing pre-arrival history reduces it by 1.97 points; and zeroing retrieval reduces it by 0.99 points. Genuine retrieval outperforms random-neighbor retrieval by 0.97 macro-average points, although a paired task-level sign test does not establish a uniform advantage (p=0.067). On the external MC-MED cohort, supervised fine-tuning reaches 0.8611 AUROC, compared with 0.8227 for locally trained gradient boosting and 0.7841 for source-only PRISM. These results support structure-aware multimodal integration and supervised external adaptation.

Agentic orchestration delegates tasks to sub-agents with independent contexts, providing a key path to extending the effective execution horizon of large language model agents. However, existing orchestration methods face two limitations: in execution, current methods fix sub-agents' goals and harnesses at creation time, limiting their adaptability to dynamic real-world environments; in optimization, sparse and delayed feedback constrains effective learning of the orchestration policy. To address these limitations, we introduce POrchestra, an on-the-fly proactive agentic orchestration framework. It allows sub-agents to proactively communicate with the orchestrator at any logical point to request harness adjustments while maintaining continuity of local execution. This mechanism not only enables sub-agents to adapt seamlessly to dynamic environments but also naturally inspires On-Policy Orchestration Distillation (OPOD) based on hindsight feedback. OPOD attributes sub-agents' proactive revision requests back to preceding orchestration decisions to derive targeted improvement signals, and leverages these signals as process supervision alongside the final reward to optimize the orchestration policy. We evaluate POrchestra on three long-horizon agent benchmarks spanning deep research, dynamic digital assistance, and software engineering. POrchestra improves task success by an average of 9.1% relative to the strongest baseline; combined with OPOD, it improves small-orchestrator success by an average of 10.0% relative to existing training baselines. POrchestra also achieves a more favorable cost-performance trade-off. Our code is publicly available at https://github.com/XueyangFeng/POrchestra.

Representational similarity measures (RSA, CKA, Procrustes) are chosen for the transformations they treat as irrelevant, and all of them take the set of conditions as given. That set is fixed by two decisions, which inputs are admitted and how they are grouped, both taken before the matrix a measure operates on exists: no invariance property reaches them. We measure what they are worth where the arrangement the conditions ought to have is known independently of the systems compared, the acoustic organisation of vowels, across fifteen languages and twenty-seven representations. Changing which categories are compared, at a fixed number of them, moves the gap between two representations by more than twice the effect being measured, while changing how many moves it not at all. In this setting rankings of models are identified under these choices, and rankings across languages are not.

Cyber-Physical Systems are developed through multiple model based activities, comprising several stages of development. As the development progresses through simulation, experimentation, and operation, observations of system execution may reveal missing, uncertain, or invalid design knowledge. The path from execution results to persistent design-time enrichment, however, is often implicit. This paper addresses this challenge by proposing an evidence driven workflow for systematically transferring knowledge produced during simulation and operation back into design-time models. We demonstrate the workflow using an Influence Model, a design-time artefact used to represent how design artefacts and environmental factors jointly affect system response properties, using a sign-following robot use case. We show how simulation evidence can enrich Influence Models by supporting participant pruning decisions, interaction characterisation, influence-function abstraction, and validity-scope information. The application demonstrates the feasibility of the workflow for simulation-based evidence and its potential benefits to improve design-time reasoning and feedback in subsequent development iterations.

Speech recognition on low-cost microcontrollers is, in practice, limited to keyword spotting or a fixed list of commands. We present Oído, an inference engine that runs open-vocabulary English speech recognition entirely on an ESP32-S3 (240 MHz dual-core Xtensa LX7, 512 KB of SRAM, 8 MB of PSRAM, 16 MB of flash, no neural accelerator). Oído executes a 13M-parameter Conformer-CTC model with int8 weights, per-frame dynamic int8 activations and int8 attention on the chip's SIMD instructions, with a schedule designed around flash and PSRAM bandwidth. On LibriSpeech test-clean/test-other the on-device arithmetic reaches 3.70%/8.23% word error rate (WER), within 0.12 points of full precision, and 3.33%/7.15% with an on-chip 1.3M-parameter language model; a transducer head gives 2.99%/6.67% and an int4 variant fits in 8.3 MB. This is 2.3-2.8x fewer errors than the vendor's recognizer on the same chip, and fewer than Whisper tiny.en and Moonshine tiny in full precision on a laptop, including in real noise and reverberation. Measured on a physical board, the firmware runs at 1.97x the duration of the audio on one core, with transcripts identical to an instruction-exact emulator: accurate, but not yet real time (a 4 s command is transcribed about 8.7 s after it ends). The two-core schedule gives wrong results on silicon and is disabled. A streaming variant shows partial text while the speaker talks and delivers the final text 4.1-11.1 s after clips of 3.6-10.6 s (7.2-22.4 s in utterance mode) at 4.86%/11.03% WER; a Spanish model makes 34-60% fewer errors than multilingual Whisper tiny. Code and models are released.

Recent accent classification systems can achieve high accuracy in standard in-corpus evaluations. However, when speakers overlap between the training and test sets, such results may overestimate model robustness compared with real-world deployment scenarios. This study revisits the robustness of DistilHuBERT-based accent classification by using L2-ARCTIC and multiple external English accent datasets. The comparison covers file-level and speaker-disjoint splits, along with balanced external-data training and teacher-student pseudo-labeling. The large performance gap between in-corpus and external evaluations suggests that file-level evaluation substantially overestimates performance, while speaker-disjoint evaluation can provide a more realistic estimate of the generalization to unseen speakers. In this study, models trained on external data alone do not reliably generalize to L2-ARCTIC. Among the tested configurations, balanced external-data training offers the most promising supervised setting. The pseudo-labeling experiments further indicate that confidence-based filtering alone is insufficient under domain mismatch. Overall, this study suggests that robust accent classification evaluation should simultaneously include speaker-disjoint evaluation and cross-corpus testing.

Earnings call transcription presents a realistic continual learning scenario where Automatic Speech Recognition (ASR) systems are updated with new recordings but risk catastrophic forgetting of previously learned financial speech patterns. Adapting Whisper-small from a general-domain speech corpus to the Earnings22 benchmark reveals the opposite: all evaluated methods exhibit improvement on held-out financial speech from the same domain during adaptation rather than forgetting. This task complementarity makes regularization counterproductive. Nevertheless, when regularization is warranted, a layer-wise mini-Kronecker-Factored Approximate Curvature (mini-KFAC) approximation seems to achieve lower word error rates (WERs) and more stable optimization than standard diagonal Elastic Weight Consolidation (EWC). The results suggest that task complementarity should be assessed before applying continual learning constraints, and prefer mini-KFAC EWC over diagonal EWC when needed.

Tibetan, spoken by some seven million people on and around the Qinghai-Xizang Plateau, remains underrepresented in natural language processing (NLP). We present, to our knowledge, the first peer-reviewed systematic bibliometric survey of Tibetan NLP: 140 publications from 2004 to 2026 in the ACL Anthology and DBLP, annotated for venue type, research area, first-affiliation country, and resource release. Output has grown rapidly, with 84% of the publications appearing since 2018, and has expanded bottom-up along the NLP pipeline, from corpora and word segmentation through script, speech, and applications to large language models (LLMs). Researchers based in China first-authored 87% of the publications, which span all 25 represented research areas, whereas work from other countries concentrates on Classical Tibetan. We argue that this full-pipeline paradigm, realized under the compounded constraints of a complex script, dialectal diversity, and geographic remoteness, offers a reference model for other low-resource languages of the Global South. We also identify gaps in dialect coverage, responsible NLP, native and comparable evaluation, cross-community visibility, and resource release (only 29% of the publications release data, code, or models), and recommend open science, shared standards, and community participation for the LLM era.

Coding agents are increasingly used to build programs from scratch as software engineers do rather than to resolve issues within a repository or to generate a repository from a given requirements document. However, scalable verifiable training environments that force agents to discover what is required and to decide how the system should be designed remain scarce. Moreover, existing agent workflows also cannot sustain long-horizon development in which the agent explores the reference program before implementation. To address these challenges, we introduce ProgramGym, a framework consisting of Program-Env and Program-Agent for scalable training environments of long-horizon program reproduction agents. Specifically, Program-Env constructs verifiable training environments from repositories through a feedback-driven loop that repairs runtime sandbox builds and the quality of black-box tests. Using this novel pipeline, we obtain 1,648 training environments spanning C, C++, Go, and Rust. Program-Agent combines a role-isolated workflow of analysis, implementation, and review with revision management for compilation checks and failure recovery. Experiments on ProgramBench show that both Program-Env and Program-Agent improve performance on program reproduction. Fine-tuning on trajectories with these constructed environments raises the performance of Qwen3.8-27B from 34.87% to 42.84%. Program-Agent further raises the base and SFT models to 55.71% and 56.25%, both exceeding some flagship models. We will release the ProgramGym to support further research on coding agents that build programs from scratch.

Stance detection remains challenging for large language models (LLMs) due to implicit opinions, sarcasm, and context-dependent expressions of stance. Although automatic prompt optimization offers an alternative to fine-tuning without updating model parameters, existing structured approaches are not specifically designed around stance-label semantics and class-specific ambiguities. We introduce MoPrO-SD, a stance-specific prompt optimization framework that decomposes prompts into functional semantic modules and refines them sequentially using LLM-based feedback. We study three refinement strategies: instruction-only optimization (I), instruction with reasoning (I+R), and instruction with per-example reasoning (I+R+E), using class-conditioned errors for stance-label refinement. We also introduce VaccineStance, a new human-annotated benchmark for vaccine stance detection. Experiments across four datasets and five open-weight LLMs show that MoPrO-SD improves upon the human-written Base prompt in many settings. On SemEval-2016, class-conditioned feedback achieves higher mean Macro-F1 than pooled feedback for four of five backbones. Both reasoning-based strategies, I+R and I+R+E, achieve strong performance, with their relative effectiveness varying across models and datasets, while remaining competitive with recent prompt optimization methods. The MoPrO-SD code and VaccineStance dataset are available on GitHub.

Rubrics have been widely used to evaluate and optimize language models on open-ended tasks. However, existing work primarily focuses on generating better rubrics, with limited understanding of what makes rubric design effective. In this work, we conduct a systematic mechanistic analysis of rubric design across criterion organization, criterion operationalization, and score aggregation. Our analysis shows that structural organization enables more comprehensive criterion construction, semantic specification alone does not ensure effective criterion-level judgments, and sophisticated score aggregation provides limited additional gains. Guided by these findings, we introduce Adaptive Rubric Tree Search with Hybrid refinement (ARTS-Hybrid), which combines adaptive tree construction with semantic-first, response-calibrated refinement. Experiments across benchmarks show that ARTS-Hybrid improves both end-to-end evaluation performance and intrinsic rubric quality. Further experiments suggest that, beyond evaluation, rubrics generated by ARTS-Hybrid provide effective signals for model optimization.

Mixture-of-Experts (MoE) models have become the standard architecture for frontier LLMs. The multiple parallel experts in each transformer block theoretically enable expert specialization according to categories such as domains, languages, or modalities during pretraining. However, when data is imbalanced across categories, minority categories can become concentrated in a small fraction of the expert pool or spread inefficiently across them, effectively limiting their available capacity. In this work, we demonstrate that category-conditioned expert-selection biases can preferentially route tokens from different categories toward specialized expert pools. By applying biases based on input category, we provide a simple way to induce soft category-aligned specialization. This intervention encourages categories toward distinct expert subsets while retaining cross-category sharing when useful. More importantly, it enables the over-allocation of expert capacity to categories that would otherwise be disadvantaged by data scarcity. We observe performance improvements across modalities, low-resource languages, and code in models up to 10.6B parameters. At our smallest and largest model scales, these gains are equivalent to allocating 18.5% and 2.7% additional baseline training tokens, respectively. Overall, our results show that expert biases can serve as a lightweight control for category-level expert allocation, making room for low-resource categories while retaining standard MoE inference.