Papers

Two clinical AI systems can score nearly identically on coverage-based rubrics yet behave radically differently when their patient inputs change: one updates its recommendations to match the new clinical signal, while the other produces the same output regardless. We introduce the Causal Sensitivity Score (CSS), a pre-registered interventional metric that mutates oncology tumor-board cases along five clinically meaningful dimensions - biomarker flips, prior-treatment failures, biomarker removals, surgery-status changes, and stage perturbations - and scores whether each model updates its recommendations in the pre-registered correct direction using a {0, 0.5, 1.0} scale. Benchmarked against the Consensus Match Score (CMS), a coverage-based weighted recall metric, six frontier models from three labs evaluated in single-shot inference across 224 cases rank in nearly opposite orders: all six models change rank, the CMS-worst model becomes CSS-best, and one upper-mid CMS model ranks last on CSS. We further surface a universal safety blind spot: every frontier model fails on surgery-status interventions (at most 17.2% CSS on Family D), a finding CMS does not expose. The metric also transfers to tool-using agents: in a ReAct-style experiment, tool use improves CSS for five of six models (+2.5 to +20.3 percentage points), yet the lowest-CSS model retrieves the same chart sections and still fails to update its recommendations - revealing a structural responsiveness deficit visible only under counterfactual evaluation. Cross-judge replication and three-rater medical-professional validation confirm the aggregate findings. Interventional pre-registered metrics like CSS complement coverage-based evaluation for clinical AI agents: they capture responsiveness that coverage metrics miss and offer a candidate dense reward signal for future agentic RL systems.

U.S. immigration law spans thousands of pages of official policy, federal regulations, and procedural guidance that change frequently and carry high stakes for petitioners who lack legal representation. We describe the construction of ImmigrationQA, a source-grounded question-answering dataset of 17,058 pairs across 13 immigration subdomains, and the fine-tuning of a Llama 3.2 3B Instruct model on that dataset using parameter-efficient LoRA. The corpus was assembled from 11 primary and secondary sources – including the USCIS Policy Manual, 8 CFR, BIA precedent decisions, and community Q&A – yielding 10,056 validated canonical documents and 18,308 text chunks. Structured QA pairs were generated from these chunks using Claude Sonnet 4.6 via five mode-specific prompts, with 22 pairs rejected for insufficient source-span overlap. The fine-tuned model was evaluated against a held-out split of 993 pairs using LLM-as-judge scoring on a 101-example stratified sample. The fine-tuned model scored a mean of 1.08/3.0 (16.8% fully correct; 101-example stratified eval) versus the Llama 3 8B base model at 0.85/3.0 (4% fully correct), a relative improvement of 27% in mean score; a zero-shot Claude Sonnet baseline scored 1.52/3.0 (25% fully correct). The fine-tuned model shows concentrated improvement in procedural subdomains (travel documents, adjustment of status, nonimmigrant visas) while remaining weak on complex legal reasoning and time-sensitive statistics. The full pipeline ran for approximately $29 in cloud compute. All artifacts – dataset, model, code, and prompt templates – are publicly released. The system is not a substitute for legal counsel and does not reflect regulatory changes after the corpus crawl date.

While platforms like Google Scholar and Semantic Scholar track citations for academic papers, no comparable infrastructure exists for monitoring dataset usage in research literature, leaving the landscape of data use largely opaque. Addressing this gap is critical for transparency, reproducibility, and monitoring of impact, yet progress is hindered by inconsistent citation practices, scarce labeled data, and ambiguous references to datasets in the wild. Traditional NLP approaches struggle with these challenges, motivating the shift toward more adaptive, semantically rich models. Building on prior work using LLMs for data mention detection and synthetic data for bootstrapping training, this paper presents an updated methodology for scalable dataset monitoring. We introduce a multitask GLiNER-based framework that jointly performs dataset mention extraction, relation identification, and usage-context classification. To address label scarcity, the pipeline leverages synthetic data generation to produce training examples and LLM-based revalidation to filter incorrect mentions and enforce labeling consistency, together improving reliability, coverage, and output consistency across the training pipeline. This work advances the development of open-source tools for monitoring data use in research literature, contributing to the broader goal of generalizable, unconstrained dataset citation tracking.

Speculative decoding is a popular technique for large language model (LLM) inference, enabling faster generation by drafting multiple tokens with a smaller draft model. However, the effectiveness of speculative decoding has mainly been studied for English. Motivated by the curse of multilinguality, we hypothesize that speculative decoding is far less effective for low-resource languages due to the limited multilingual capacities of smaller models. We test eleven languages under a standard speculative decoding setup and find strong evidence for our hypothesis. Next, we try to improve the multilingual capabilities of the smaller draft model via distillation from the larger model. We find, though, that distillation generalizes poorly across tasks in the same language, and we argue that assembling a task-agnostic, fully representative dataset is infeasible for low-resource languages. Finally, we propose weaker n-gram models as draft models; these provide moderate speed-ups due to their minuscule inference cost.

Prior KV cache compression schemes empirically demonstrate that the prompt cache is partially redundant during decoding, dropping or summarising entries with little accuracy loss. We ask when and what kind of redundancy: at which layers, after how many decoding steps, and in what form can the prompt span KV cache be replaced without breaking the task. A controlled splice intervention swept over layer cutoff and decoding steps shows this redundancy is about form (chat template scaffolding) rather than content. Replacing the upper layer prompt span KV cache with KV cache from a chat template scaffold whose user content is a neutral filler recovers near clean accuracy, while zeroing the same slots collapses accuracy. The dissociation replicates across the Qwen3, Gemma 3, and Llama 3 families on multiple datasets.

LLM-as-a-Judge is a scalable alternative to human evaluation, yet existing rubric-based methods rely on human-annotated data such as reference answers or expert-crafted rubrics. We propose to automatically generate fine-grained evaluation rubrics without any human annotation. Our training-free method generates rubrics at dataset-specific and instance-specific granularities, achieving performance competitive with existing methods across four benchmarks. We further present a method that iteratively fine-tunes a rubric generator model via meta-judge reward signals. The fine-tuned generator outperforms all existing baselines in both pairwise and pointwise evaluation. Notably, a fine-tuned 14B rubric generator outperforms a much larger proprietary model at rubric generation, showing the effectiveness of our fine-tuning strategy.

Spatial reasoning benchmarks typically evaluate whether vision-language models can derive the correct answer from a visual observation. Yet in real 3D environments, the observation itself may be unreliable: occlusion can remove task-relevant evidence, while perspective can make visible geometry misleading. Reliable spatial reasoning therefore requires more than answering a question correctly. A model must also assess whether its current observation provides sufficient and trustworthy evidence for that answer. We introduce SPATIALUNCERTAIN, a controlled evaluation framework for studying viewpoint-dependent observational uncertainty. We study two complementary failure modes: missing evidence caused by occlusion and misleading evidence caused by perspective. We further evaluate whether models can recognize when the current view is unreliable and identify a more informative observation. Across eight open- and closed-source vision-language models, we find that model behavior does not track the reliability of visual evidence. Models do not reliably become more cautious as evidence disappears, and under perspective conflict, their judgments increasingly follow projected appearance rather than the unchanged physical 3D relation. Internal analysis suggests a corresponding representational asymmetry: projected 2D relations are readily available, whereas the underlying physical 3D relation is barely decodable. Moreover, models that can identify an informative viewpoint when explicitly asked often fail to recognize when such an additional view is needed. These failures are not fully resolved by prompting or fine-tuning, and providing a better viewpoint is substantially more effective than adding depth information to the same misleading observation. Our results identify assessing the reliability of visual observations as a distinct and missing component of current spatial reasoning evaluation.

Korean grammatical error correction (K-GEC) presents a structural mismatch between word-based evaluation and the morpheme-level locus of many learner errors. Postpositions and verbal endings are bound to lexical hosts, but they encode grammatical relations that must be represented in correction and evaluation. This paper refines word-based grammatical error annotation for L2 Korean by addressing three connected problems in existing resources: surface target realization, Korean-specific edit annotation, and single-reference evaluation. We reconstruct target sentences from the National Institute of Korean Language (NIKL) L2 corpus under morphologically constrained realization rules and convert its morpheme-level annotations into word-level m2 edits. We then define a Korean ERRANT-style annotation scheme that preserves the MRU core while distinguishing functional morpheme errors, spelling errors, word boundary errors, and word order errors. We also augment the KoLLA corpus with an additional reference correction, yielding a multi-reference evaluation setting for Korean GEC. Empirical validation shows that the refined NIKL targets yield lower perplexity, the converted m2 files achieve higher agreement with source-target edit representations, and the refined resources improve KoBART-based correction under the same model setting. Multi-reference KoLLA evaluation further reduces the penalty imposed on valid corrections that diverge from a single reference, especially for neural and prompted GEC systems. These results show that Korean GEC evaluation depends not only on correction models, but also on reference data and edit annotations that reflect Korean morphology, spacing, and correction variability.

Sentence-embedding models for semantic search are overwhelmingly developed and evaluated on English corpora. When applied to clinical retrieval in other languages – particularly retrieval of ICD-10-CM / CIE-10 codes – recall degrades in ways often masked by aggregate benchmarks. We study whether large generative language models can serve as data factories to close this gap. We build a two-stage retriever (bi-encoder followed by cross-encoder reranker), fine-tuned from a Spanish biomedical encoder (PlanTL-GOB-ES/bsc-bio-ehr-es) on Gemini-generated synthetic data covering English, Spanish, Catalan, Italian, Portuguese and French, and evaluate against BioBERT-ST and the un-tuned Spanish encoder. The bi-encoder alone matches BioBERT-ST on MRR (0.876 vs. 0.866) and overtakes it on R@3 (0.650 vs. 0.626) and R@5 (0.804 vs. 0.790) without English biomedical pretraining. Adding a cross-encoder reranker lifts aggregate R@5 to 0.822 and dominates on four of five languages (+0.017 Spanish, +0.033 Catalan, +0.018 French, +0.037 Portuguese) at the cost of a small English regression. The trade-off is clinically acceptable: Portuguese reaches R@5 = 0.829 vs. BioBERT-ST's 0.714. Contributions: an open recipe for building domain-specific medical retrievers from LLM-generated data; quantification of the learning gain (MRR 0.755 to 0.876, +15.9% with ~19,500 synthetic pairs); and a characterisation of where gains concentrate by language and rank.

Post-trained language models often exhibit a recognizable AI-like writing style, yet its connection to post-training and internal representations remains poorly understood. In this work, we study whether post-training introduces or amplifies AI-like stylistic regularities and whether these regularities have a localized internal signature. To this end, we compare human text, base-model generations, and aligned-model generations under matched human-source prefixes. Aligned generations show lower human-corpus affinity and higher AI-detection rates than base generations, suggesting that post-training shifts generated text away from human-corpus style and toward detector-visible AI-like text. We then introduce PASTA (Post-training Alignment Signature Targeted Ablation), a training-free method that estimates a post-training alignment signature from aligned-base residual contrasts and ablates the corresponding direction during decoding. Across 11 aligned models and 6 AI-detectors, PASTA lowers the detection rate for most aligned models; this effect transfers well across detectors and is not reproduced by random directions. Qualitative analysis suggests that PASTA generations remain relevant and coherent while exhibiting greater stylistic variation. Together, these results show that AI-like stylistic effects of post-training can be measured, localized, and causally tested through activation ablation.

Recent work describes what transformers can and cannot compute through connections to boolean circuits, but existing results lack exact characterizations and are sensitive to modeling choices. Padded transformers – to whose input filler symbols such as “...” are appended – emerge as a useful gadget for establishing equivalences to circuit classes by providing polynomial space for adaptive parallel computation. However, only a limited set of padded transformer idealizations has been studied, leaving open how robustly these equivalences hold under changes to attention type, model width, and uniformity. We find that, under practical assumptions, padded transformers are surprisingly robust to all of these, and identify numeric precision and model depth as the main factors affecting expressivity. Concretely, we prove that polynomially padded $\text{L-uniform}$ constant-precision transformers are equivalent to $\text{L-uniform AC}^0$, while growing-precision ones achieve $\text{L-uniform TC}^0$ regardless of width. Furthermore, looping enables sequential processing analogous to circuits: $\log^d N$-looped constant-precision transformers reach $\text{FO-uniform AC}^d$, and growing-precision ones reach $\text{FO-uniform TC}^d$. Interestingly, growing width or precision beyond logarithmic does not increase expressivity, and all our results hold for both softmax and average hard attention transformers.

Large language models must frequently process untrusted inputs, such as judging an answer from another model or running tasks like spam and harm classifiers while under adversarial pressure. These inputs are often string-formatted directly into a prompt template, leaving systems fragile to manipulation. Current LLM specs from major providers like OpenAI distinguish trustworthiness along an Instruction Hierarchy, from System messages (most trusted) to Tool Results (least trusted). A possible natural mitigation is to wrap untrusted content in a mock tool call as a quarantine. We explore this hypothesis with an automated redteaming search over static attack strings across seven models and three LLM-as-a-Judge tasks. Counter to our hypothesis, tool-wrapping does not broadly improve robustness. On a binary evaluation task (GSM8K grading) it typically increases attack success rates, an apparent inversion of the instruction hierarchy. On scalar and pairwise tasks the effect is smaller and model-dependent, with no tested model reliably helped, and several showing inversion. We recommend evaluating this limitation in deployed systems, and longer-term, pursuing stronger Instruction Hierarchy training or new untrusted-input primitives.

Machine unlearning evaluation is structurally skewed: Why-type questions, which probe causal and relational knowledge, comprise less than 0.06% of CounterFact, 0.6% of ZSRE, and less than 1.3% of TOFU, MUSE, and WMDP-Cyber. This near-zero representation means that methods that fail on causal knowledge can score highly in aggregate, and this failure is undetectable without balanced evaluation. We present 5WBENCH, a balanced 5,000-sample benchmark with 1,000 examples per 5W category (Who, What, When, Where, Why), making causal unlearning failures quantifiable for the first time. Using 5WBENCH, we show that no existing baseline simultaneously achieves high forgetting and high retention on Why-type questions: aggressive forgetting degrades retained knowledge, while conservative methods fail to forget causal facts. Why-type difficulty stems from multi-hop reasoning chains (44% of Why entries vs. less than or equal to 2% for others) and gradient dilution over 40.1-token answer spans. We present MAAT (Multi-phase Adapter-Aware Targeted Unlearning), a three-phase framework operating on LoRA adapter weights, combining gradient-projected ascent, SVD rank-dimension pruning, task vector negation, and hybrid KL-hidden-state retain repair. MAAT is the first method to simultaneously achieve high forgetting and high retention on Why-type causal knowledge, reaching a new operating point on the forget-retain Pareto frontier. We make our code publicly available.

LLM benchmark labels are frozen at release and silently propagated into downstream benchmarks, errors and all. We introduce an Item Response Theory-based indicator that surfaces likely mislabels at 95% precision in the top 200 examples across seven preference and multiple-choice benchmarks using responses from 114 models, outperforming a supervised classifier. We trace these errors to mechanical labeling heuristics, upstream annotation mistakes inherited unchanged from source datasets, and fundamentally ambiguous items without a defensible single label. The same model fit reveals that reward models specialize in stylistic preference rather than factual knowledge, and identifies one frontier reward model that agrees with detected mislabels at 78% accuracy versus 38% for its peers, consistent with benchmark contamination or benchmark-specific over-optimization.

Watermarking embeds statistical signatures in AI-generated text for detection and attribution. We reveal a fundamental vulnerability: when users access multiple models (today's reality), watermarks trivially fail. Watermarks perturb output distributions away from the original, and in competitive markets, these perturbations are typically independent across providers. We theoretically prove that averaging output probability distributions recovers the unwatermarked distribution with up to a second-order error term. Empirically, simply averaging 3-5 models cancels out these perturbations. We introduce WASH (Watermark Attenuation via Statistical Hybridisation), which solves practical challenges in ensemble generation: vocabulary misalignment and tokenisation differences across heterogeneous models. Experiments across six watermarking schemes and three LLMs show that averaging across 3 models suppresses detection z-scores from 5-300 to below 2 (below the detection threshold of 4) and reduces TPR at 5% FPR to below 50%, while improving quality by 27.5% and running 6 times faster than the best baseline on the long sequence generation. Our results suggest that robust AI-text detection via watermarking requires either accepting this fundamental vulnerability or unprecedented coordination among model providers.

RAG-based legal assistants have been growing in popularity, but LLM hallucinations remain a key issue and potentially undermines justice. While benchmarks have been developed to evaluate progress, many rely on synthetic queries rather than realistic legal scenarios. Moreover, Canadian law remains underrepresented in existing evaluations. To address this gap, we introduce CanLegalRAGBench, a Canadian legal QA benchmark based on realistic queries and expert-annotated answers grounded in case law. Our evaluation shows that retrieval performance is sensitive to design choices and that open-source embedding models are competitive with closed source models. However, it also reveals the limitation of automatic evaluations that penalize systems for retrieving alternative relevant documents. We also find that generated answers often diverge from gold responses, either with hallucinations or by producing overly detailed or irrelevant content, with 8-29% of claims not being supported by the retrieved documents. We hope this benchmark will help drive continued progress in addressing limitations of legal RAG systems.

Aligning large language models (LLMs) to heterogeneous and rapidly evolving safety requirements remains a critical challenge. Existing instruction-tuned LLMs and standalone safety classifiers often fail to generalize to new safety configurations, motivating the need for Reward Models (RMs) that are explicitly configurable to changing specifications. We introduce the Configurable Safety Reward Model (CSRM), which is jointly optimized for calibrated safety compliance and reward modeling. Our approach is supported by configuration-targeted data augmentation that enforces instruction adherence while preserving relative severity structure. The resulting RM is sensitive to fine-grained safety configurations and conversational nuances, substantially improving generalization to previously unseen safety configurations. CSRM achieves state-of-the-art performance on recent configurable safety benchmarks, including CoSApien (94.6% F1) and DynaBench (75.8% F1), without requiring additional human annotation. When used for downstream safety alignment, CSRM yields LLMs with a significantly improved helpfulness-safety tradeoff compared to existing baselines.

Large language models (LLMs) are widely used as cross-lingual knowledge interfaces. However, culturally grounded questions often reflect globally dominant narratives rather than local contexts. We study this failure mode as global narrative dominance in Bangla, a low-resource cultural context. We introduce CulturalNB, a dataset of 717 manually curated Bengali cultural instances with parallel Bangla–English question–answer pairs and supporting evidence, metadata, and sociocultural annotations. Using question-only and evidence-based prompting, we evaluate nine state-of-the-art LLMs with human and two independent LLM judges across metrics for cross-lingual consistency, language anchoring, global substitution, institutional bias, and epistemic perspective coverage. Results show that questions asked in English systematically increase global substitution and institutional framing while reducing local perspective coverage. Local evidence improves factual consistency and perspective coverage, but does not eliminate language-induced epistemic shifts. These findings suggest that cultural failures in LLMs are not only missing-knowledge errors but also failures of grounding and narrative prioritization.

Reinforcement learning with verifiable rewards (RLVR) trains language models using programmatically checkable signals such as unit-test outcomes, enabling direct optimization for functional correctness in code generation. We conduct an empirical study of RLVR for Python code generation on the MBPP benchmark using two small models (Qwen3-0.6B and Llama3.2-1B) with LoRA fine-tuning. Across multiple reward formulations such as: unit-test-only rewards, static-analysis-only shaping via the Ruff linter, and a combined reward, we compare group-based policy optimization variants (GRPO and GSPO) and evaluate both functional correctness and behavioral diagnostics. In our experimental setting, RLVR improves pass@1 on MBPP test by up to 13 percentage points under proposed combined reward configuration. However, we find that reward shaping can induce systematic behavioral shifts: using only static-analysis penalties may bias the policy toward shorter completions that reduce lint errors without reliably improving functional correctness. In contrast, combined rewards mitigate this degeneration and yield more stable trade-offs between correctness and style constraints. Overall, our results highlight that RLVR effectiveness for code generation is highly sensitive to reward design and optimization granularity, and that diagnostics beyond pass@1, including generation length, Ruff severity profiles, and execution error types are useful for identifying failure modes.

As large neural models have become better at language tasks, researchers are increasingly building multi- and omnimodal models that handle more modalities of data. One example is the expansion of speech recognition models to audio-visual data for noise mitigation and multimodal subtitling. While performance and bias have been studied extensively in the single-modality regime, it is unknown how new modalities affect this, even though they produce biases in humans. We therefore propose the first bias evaluation of multimodal speech recognition, where we create videos pairing different faces with the same audio, and measure changes in speech transcription accuracy. We find large quality-of-service differences across mWhisper-Flamingo and Gemini models, with drops of up to 4.05 word error rate points, across self-declared gender, ethnicity, and their intersection. Our findings point to a priority for developers to evaluate, fix, and communicate such limitations, as providing more signals through additional modalities is not necessarily better, and may even lead to biased outcomes.

Multi-label topic classification without labeled training data is a challenging task, specially when documents contain complex relational information. We present a zero-shot multi-label topic classification framework and systematically investigate how per-article knowledge graph augmentation affects its performance. The base framework classifies topics in documents without labeled training data and has four variants: article-only classification, keyword-enhanced classification, and self-consistency decoding variants of both. Then, we augment each base variant with per article knowledge graph. This graph is extracted from the input document through a pipeline similar to KGGen based on subject-predicate-object triples. We test all eight methods, four base and four graph augmented on fifteen LLMs and eight multi-label datasets across different domains. For the base framework, keyword-enhanced classification (AK) is the best performing method, and six out of fifteen LLMs surpass the sentence-encoder baseline. Graph augmentation has positive and negative impacts on small and large models, respectively. This shows that larger models already contain enough relational information from pretraining. Furthermore, the self-consistency decoding variant does not show performance improvements in any experiment while increasing computation costs about fivefold.

Large language models (LLMs) remain limited on tasks requiring indirect reasoning, cultural knowledge, and coordinated hypothesis testing. We investigate whether team-based interaction improves LLM performance in What? Where? When? (ChGK), a quiz game designed to reward collective reasoning. We introduce three team strategies: Voting, Silent Team (the captain observes final answers), and Talkative Team (the captain observes both answers and rationales). To minimize data leakage, we evaluate these strategies on a dataset consisting of 572 ChGK questions released in 2025. Using six recent large-scale open models, we show that team-based strategies outperform single-model baselines, yielding gains of up to 20 percentage points in accuracy. The best team achieves 44.23% accuracy, and approaches human team performance on questions with available human statistics. Analysis of inter-model diversity reveals that disagreement strongly predicts lower accuracy, but explanatory communication substantially mitigates performance drops. We further examine captain behavior and find no evidence of self-preference bias; access to peer rationales improves captain judgments. Overall, LLM teams function primarily as answer selection and error-filtering mechanisms rather than generators of novel solutions. Our findings highlight the importance of interaction and suggest adaptive strategies as a promising direction for multi-agent systems.

Regional accent classification in Brazilian Portuguese (pt-BR) suffers from the need for reliable labeling. While large self-supervised learning (SSL) speech models are powerful, their training pipelines dilute sociophonetic information, since accent labels are generally not reliable or are not used in training objectives. This work introduces a novel workflow for feature extraction using only acoustic labels. By isolating explicit regional accent landmarks and using a phoneme-based forced aligner (ZIPA), our targeted feature set captures dialectal variance more effectively than utterance embeddings, demonstrating that localized features can outperform general-purpose architectures on accent-related tasks using minimal and objective data labels.

Black-box LLM distillation is usually evaluated as an output-matching problem: a student is considered successful when its responses are semantically similar to, or task-consistent with, those of a teacher. However, output similarity does not imply that the student is behaviorally indistinguishable from the model it imitates. We introduce bounded behavioral indistinguishability, formalized as $(\epsilon,q,t,\mathbb{A})$-behavioral indistinguishability over an explicit prompt distribution, where $\epsilon$ bounds distinguishing advantage, $q$ bounds oracle queries, $t$ bounds computation, and $\mathbb{A}$ denotes the adversary class. We instantiate this notion on Qwen and Llama teacher-student pairs using a controlled $5,000$-prompt behavioral probe suite. For each family, we compare the teacher with both the base student and the LoRA-distilled student, measuring whether distillation reduces distinguishability rather than merely improving similarity. LoRA raises semantic similarity from $0.788$ to $0.862$ for Qwen and from $0.814$ to $0.874$ for Llama. Yet adversarial evaluation reveals remaining behavioral differences: learned discriminators retain nonzero advantage, and pairwise category analysis shows artifacts concentrated in style/format, robustness, and domain-technical prompts. A pairwise teacher-identification adversary confirms this trend. With a different-family Llama judge and A/B-swap consistency filtering, Qwen distinguishing advantage drops from $0.158$ for the base student to $0.081$ after LoRA distillation. Query-budget experiments show that disagreement-guided acquisition does not consistently outperform stratified random sampling, indicating that coverage and diversity remain strong baselines. Our results show that semantic fidelity is useful but insufficient: black-box LLM distillation requires bounded, adversarial, and category-aware evaluation.

Cross-lingual activation steering has been studied mainly for language-agnostic properties such as factual knowledge and reasoning. Its transferability remains unclear for figurative language, whose expressions are often tied to a particular language. In this work, we study cross-lingual steering of figurative language: whether steering vectors learned in one language can affect generation in another. We first show that within a single language, steering effectively increases the ratio of the intended figure of speech, especially for metaphor and idiom. Then under the cross-lingual setting, across 408 comparisons spanning four multilingual models, six languages, and four figures of speech, steering increases the figurative ratio from the unsteered 12.1% to 20.4%. To understand the mechanism, we decompose each language's own steering vector into the mean of steering vectors from other languages and a residual, and show that this mean direction matches or outperforms the language's own vector while the residual steers worse. This suggests that, despite the highly language-dependent nature of figurative expressions, a partially shared direction exists in the representation space that enables figurative knowledge to transfer across languages.

Real-world data analysis is inherently iterative, yet existing benchmarks mostly evaluate isolated or short interactive tasks, leaving agents' ability to track evolving analytical context over long horizons untested. We introduce LongDS, a benchmark for long-horizon, multi-turn data analysis where agents must maintain, update, restore, and compose evolving analytical states. LongDS comprises 68 tasks constructed from real-world Kaggle notebooks, spanning 2,225 turns across six domains including Geoscience, Business, and Education. Tasks are designed around state-evolution patterns (e.g., counterfactual perturbation, rollback, multi-state composition), with an average dependency span of 11.3 turns. Evaluating five state-of-the-art models, we find that the best model reaches only 48.45% average accuracy, performance drops nearly 47 points from early to late turns, and long-horizon errors account for 52%–69% of failures. Further analysis shows that additional agent steps do not necessarily improve performance, suggesting that the key bottleneck is maintaining a correct analytical state rather than increasing interaction budget. We release LongDS to support research on reliable long-horizon agentic data analysis. Code and data are released at https://github.com/zjunlp/DataMind.

We investigate how domain adaptation reshapes explanatory behavior in language models using historical cosmology as a controlled setting. In Phase 1, we train a small language model from scratch on a pre-Copernican corpus from which explicit heliocentric references were removed, and evaluate whether Earth-motion or heliocentric continuations nevertheless emerge. In Phase 2, we fine-tune a larger pretrained model using QLoRA on the same corpus in order to study how adaptation modifies explanatory framing and cosmological stance. Model outputs are evaluated using an LLM-as-judge framework that labels both cosmological stance (geocentric, heliocentric, or ambiguous) and explanatory frame (premodern versus modern). In the constrained setting of Phase 1, the smaller models occasionally generate local Earth-motion continuations, but these remain globally unstable and insufficient to support coherent cosmological reasoning. In Phase 2, fine-tuning induces a large and statistically significant shift toward premodern explanatory framing, while the conditional cosmological stance distributions remain comparatively stable within those frames. As a result, increases in geocentric outputs arise primarily from redistribution over explanatory regimes rather than from direct modification of stance. These results suggest that domain adaptation may primarily reshape the linguistic frameworks from which continuations are generated, with changes in stance emerging secondarily from those shifts.

Large Language Models (LLMs) have demonstrated strong performance on general tasks, while often struggling to adapt to specialized domains without high-quality domain-specific data. Existing LLM-based data curation methods primarily rely on human-designed workflows, leaving it unexamined whether LLMs can autonomously execute an end-to-end data engineering pipeline for model specialization. We formalize Autonomous Agentic Data Engineering, a novel task designed to evaluate LLMs as autonomous data engineers that drive model specialization through end-to-end data curation. We frame data as an optimizable component and study agents that plan, generate, and iteratively optimize training data across multiple domains, guided by post-training performance improvement. Experiments show that autonomous LLM data engineers yield substantial gains, as GPT-5.2 constructs a training curriculum that improves a student model by 57.29%, entirely through iterative, agent-driven data adaptation. By illuminating both potential and bottlenecks, our study establishes autonomous data engineering as a measurable capability and charts a path toward agent-driven model specialization (Code will be released at https://github.com/zjunlp/DataAgent).

We present a protocol to evaluate ChatGPT's ability to generate disease-centric biomedical associations. It outlines how we generate the associations, validate the biological entities using biomedical ontologies, and verify associations using literature. The protocol includes a self-consistency strategy to assess generative reliability across ChatGPT models. To address ontology exact-match limitations, we provide a use case performing semantic verification through a workflow enabled by Retrieval-Augmented Generation (RAG) powered by open-source large language models (LLMs). This enables LLMs to establish truth over content generated by other LLMs and expose hallucination.

Human reasoning has long been theorised to operate socially, not through isolated individual cognition, but through collective adversarial discourse, a framework known as the Argumentative Theory of Reasoning (ATR). Rather than relying on individual "intellectualist reasoners" as the primary vehicle for truth-seeking, ATR reconceptualises truth as an emergent property of social epistemology: the product of imperfect individual reasoning refined under the adversarial pressure of debate. This distributed method of collective intelligence has guided humanity to ever-greater epistemic heights and underpins the foundational principles of all democratic systems. This thesis breaks new ground by, for the first time, simulating ATR through the multi-agent debate (MAD) of large language models (LLMs). With rigorous empirical analysis, we demonstrate that, when correctly engineering an epistemically diverse set of models, LLM-MAD can significantly improve truth-seeking performance on questionnaire-based tasks, even when individual debate participants exhibit limited standalone performance. Furthermore, we present strong empirical evidence that this performance gain is mechanistically grounded in the central principles of ATR, suggesting that collective reasoning may be universally favourable over individualist reasoning, rather than a quirk in biology or evolution. Finally, drawing on our analysis of debate dynamics, we propose a novel benchmarking methodology that leverages LLM-MAD to measure intrinsic model properties (such as hallucination propensity) in order to compare models in ways that current static benchmarking approaches cannot support.