BanglaVerse introduces a culturally grounded benchmark evaluating vision-language models on Bengali culture across standard Bangla, four historically linked languages, and five regional dialects. Built from 1,152 manually curated images expanded to ~32.3K artifacts, the work reveals that standard Bangla evaluation substantially overestimates model capabilities compared to dialectal settings. The core finding—that missing cultural knowledge, not visual grounding alone, is the primary bottleneck—challenges conventional multimodal evaluation practices for underrepresented languages.

Narrative similarity is inherently interpretive—different valid readings can yield divergent judgments, challenging benchmarks that encode single ground truths. This paper proposes embracing multiperspectivity by ensembling 31 LLM personas, ranging from literary critics to lay characters, to predict which of two stories is more similar to an anchor. The approach leverages Condorcet Jury Theorem-like dynamics to improve accuracy, achieving 0.705 on SemEval-2026 Task 4 while revealing that diverse practitioner perspectives yield better ensemble gains despite lower individual performance.

This paper introduces Cross-Context Verification (CCV), a black-box method for detecting LLM benchmark contamination by solving the same coding problem $N$ times in isolated sessions and measuring solution diversity. The key insight is that memorized solutions are deterministic while genuine reasoning produces natural variation. The paper pairs this with Hierarchical Cross-Context Architecture (HCCA), a multi-agent analysis framework that uses strict information restriction to prevent confirmation bias. As coding benchmarks face credibility crises from solution leakage, this work targets the urgent need to distinguish reasoning from recall in SWE-bench evaluations.

This paper investigates how interrogative stances function as markers of voice and power in French-language digital news. Analyzing over 1.2 million articles from 24 outlets (2023–2024) through a mixed-methods pipeline combining LLM pseudo-labeling and qualitative annotation, the authors operationalize pragmatic concepts like answerhood and dialogicity at scale. The study reveals that questions are sparse but structurally significant, predominantly serving framing functions rather than information-seeking, and centering elite actors over diffuse publics.

Prompt2Box addresses the limitation that vector embeddings of LLM prompts conflate topical similarity with specificity, making it difficult to distinguish whether a model fails at a broad topic or only at its most constrained variants. The core idea is to embed prompts into a box embedding space where the geometric volume encodes specificity—smaller boxes indicate more constraints—and containment represents entailment relations. This geometric re-framing enables more accurate hierarchical clustering and finer-grained weakness analysis across 17 different language models.

Human annotation for subjective NLP tasks suffers from high inter-annotator disagreement. This paper introduces ReasonAlign, a protocol that exposes annotators to LLM-generated reasoning explanations (but not predicted labels) between two annotation passes. The goal is to test whether reasoning scaffolds improve annotation consistency without the anchoring bias typical of suggestion-based systems.

This paper addresses a key gap in language model research by conducting the first tightly controlled comparison between autoregressive (AR) and masked diffusion language models (MDLM). The author trains both models on identical data (50M tokens from TinyStories), identical compute budget (20K steps, batch size 32), and identical hardware (NVIDIA H100), isolating the generation paradigm as the sole variable. The work is significant because prior studies compared these paradigms at different scales or with different datasets, making it impossible to attribute observed differences to the core architectural distinction itself.

The paper tackles the challenge of enhancing long-context reasoning in Large Language Models (LLMs), a critical capability as real-world tasks grow more complex. It proposes structured table data as a solution, mathematically demonstrating via mutual information analysis that tables possess periodic non-vanishing dependencies—unlike natural language which decays polynomially—making them ideal for training long-context reasoning. The authors present TableLong, a scalable pipeline for synthesizing diverse, verifiable table data for reinforcement learning, showing significant performance gains across benchmarks.

This paper proposes SqueezeComposer, a long-form music generation framework that tackles computational constraints by applying temporal speed-up (e.g., 2×, 4×, 8×) to compress audio sequences before generation. The core idea is to generate music in an accelerated domain using diffusion models, then restore it to normal speed, theoretically enabling models to produce 10+ minute compositions with fixed memory budgets. The approach is tested on continuation, completion, and singing accompaniment tasks.

Cross-tokenizer knowledge distillation faces a fundamental alignment challenge when Teacher and Student models use different vocabularies. This paper analyzes DSKD-CMA, the state-of-the-art method for this setting, through manual chunk alignment probes and reveals that its cross-model attention mechanism captures coarse chunk structures but suffers from noisy localization with repeated tokens. Building on this insight, the authors propose DSKD-CMA-GA, which uses generative adversarial key-query matching to align distributions between models, achieving modest improvements in ROUGE-L scores that narrow the gap between cross-tokenizer and same-tokenizer distillation.

ThinkJEPA addresses the limitation of JEPA-style latent world models that rely on short, densely sampled windows, which bias predictions toward local dynamics while missing long-horizon semantics. The paper proposes a dual-temporal architecture combining a dense-frame V-JEPA branch for fine-grained motion with a sparsely sampled VLM "thinker" branch that provides semantic guidance via multi-layer feature pyramids. This matters because it attempts to marry the physical consistency of latent world models with the general knowledge of vision-language models for robust trajectory forecasting.

Weather captioning—generating natural language descriptions from meteorological time series—sits at the intersection of time-series analysis and domain-specific NLG. This paper proposes WeatherTGD, a training-free framework that treats caption refinement as gradient descent in text space: three specialized LLM agents (Statistical, Physics, Meteorology) output textual gradients that are fused via a consensus-aware mechanism and applied iteratively to improve an initial caption. The approach aims to bridge the gap between numerical forecasting and human-interpretable explanations without any model fine-tuning.

TAMTRL addresses the temporal credit assignment problem in multi-turn RL for long-context document processing. When LLMs process documents chunk-by-chunk with memory updates, standard outcome-only rewards cannot distinguish good from bad intermediate memory updates. The paper proposes using the model itself as a teacher: during training, it provides the model with filtered (relevant-only) chunks and uses the normalized token probabilities of the generated memory as turn-level rewards. This avoids expensive rollouts or external judges while providing fine-grained supervision for each turn.

This paper reframes climate disinformation detection from classification to retrieval, treating narrative core messages as queries to rank corpus texts without fixed taxonomies. They propose SpecFi, which generates hypothetical documents using community summaries from graph-based detection (NodeRAG) as few-shot examples. The approach achieves MAP 0.505 on CARDS and demonstrates robustness to high narrative variance that cripples standard baselines.

This paper investigates how users decode emotions in text-based communication through electronic nonverbal cues (eNVCs)—orthographic signals like elongation, punctuation, and emojis that approximate paralinguistic features. The authors propose a taxonomy grounded in nonverbal communication theory (kinesics and paralinguistics) and test it across three complementary studies: a content analysis developing a regex detection toolkit, a within-subjects experiment manipulating eNVC presence and sarcasm ($n=513$), and focus groups exploring interpretive strategies. The work identifies sarcasm as a critical boundary condition where eNVCs fail to aid interpretation and provides an open-source Python/R package for automated cue detection.

This paper identifies "semantic shift"—the intrinsic evolution of meaning within a text—as the root cause of embedding pathologies like anisotropy and length-induced collapse. The authors argue that pooling-based aggregation forces "semantic smoothing," where diverse sentences compromise into a diluted representation. They formalize semantic shift as the product of local evolution and global dispersion ($\mathrm{Shift}(k) = \mathrm{Local}(k) \cdot \mathrm{Disp}(k)$), showing through controlled concatenation experiments that it predicts embedding concentration and retrieval degradation better than text length alone. The work reframes geometric pathologies not as inherent model defects but as consequences of content structure interacting with pooling mechanics.

This paper investigates whether cross-lingual transfer (CLT)—prompting models to translate queries to English, reason in English, then translate answers back—can bridge the performance gap for low-resource languages. The authors benchmark eight LLMs across 2,000 responses in Kazakh and Mongolian, finding that CLT selectively benefits bilingual models (+2.2–4.3pp) but not English-first architectures, while revealing a concerning "fluency illusion" where models appear fluent in LRLs while producing less accurate content.

LLM annotations encode some human perspectives better than others, especially in subjective tasks where demographic background shapes judgments. This paper introduces Perspective-Driven Inference (PDI), a statistical framework that treats the distribution of group-specific annotations as a vector estimand $\theta^* = (\theta^*_{g_1}, \dots, \theta^*_{g_K})$ and adaptively allocates limited human labels to groups where LLM proxies are least reliable. The core contribution is an error-predictor-driven sampling rule that improves estimation accuracy for harder-to-model demographics while maintaining valid frequentist coverage.

The paper proposes the Conspiracy Frame, a semiotic and frame-semantic representation of conspiratorial narratives with five elements (plan, secret, in-group, out-group, call-to-action), and introduces Con.Fra., a span-annotated Telegram corpus. The core hypothesis is that injecting FrameNet-derived semantic frames into LLM prompts will improve conspiracy detection and explainability. Results show that while frame-guided prompting achieves comparable classification scores to few-shot learning, it does not consistently outperform it, though it reveals interesting abstract semantic patterns.

Code retrieval currently relies on dense embeddings, but this paper proposes SPLADE-Code, the first large-scale learned sparse retrieval (LSR) family for code search (600M–8B parameters). The authors address unique challenges including subword fragmentation, semantic gaps between natural language and code, and latency issues from long code documents. Their lightweight single-stage training achieves 75.4 nDCG@10 on MTEB Code under 1B parameters (state-of-the-art for that size) and 79.0 with 8B parameters, while enabling sub-millisecond retrieval via inverted indices.