# Research Papers — 2026-01 · bundle page 166 of 189

> Full-text markdown of papers #8251–8300 of 9403 published in 2026-01, newest first. AI-readable mirrors of the full 9403-paper 2026-01 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-01/page/166
[← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/165.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-01/page/167.md)

### Who Laughs with Whom? Disentangling Influential Factors in Humor Preferences across User Clusters and LLMs

Authors: Soichiro Murakami, Hidetaka Kamigaito, Hiroya Takamura, Manabu Okumura
Published: 2026-01-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.03103v2

Humor preferences vary widely across individuals and cultures, complicating the evaluation of humor using large language models (LLMs). In this study, we model heterogeneity in humor preferences in Oogiri, a Japanese creative response game, by clustering users with voting logs and estimating cluster-specific weights over interpretable preference factors using Bradley-Terry-Luce models. We elicit preference judgments from LLMs by prompting them to select the funnier response and found that user clusters exhibit distinct preference patterns and that the LLM results can resemble those of particular clusters. Finally, we demonstrate that, by persona prompting, LLM preferences can be directed toward a specific cluster. The scripts for data collection and analysis are publicly available to support reproducibility.

---

### Who Laughs with Whom? Disentangling Influential Factors in Humor Preferences across User Clusters and LLMs

Authors: Soichiro Murakami, Hidetaka Kamigaito, Hiroya Takamura, Manabu Okumura
Published: 2026-01-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.03103v1

Humor preferences vary widely across individuals and cultures, complicating the evaluation of humor using large language models (LLMs). In this study, we model heterogeneity in humor preferences in Oogiri, a Japanese creative response game, by clustering users with voting logs and estimating cluster-specific weights over interpretable preference factors using Bradley-Terry-Luce models. We elicit preference judgments from LLMs by prompting them to select the funnier response and found that user clusters exhibit distinct preference patterns and that the LLM results can resemble those of particular clusters. Finally, we demonstrate that, by persona prompting, LLM preferences can be directed toward a specific cluster. The scripts for data collection and analysis will be released to support reproducibility.

---

### TGIF: Text-Guided Layer Fusion Mitigates Hallucination in Multimodal LLMs

Authors: Chenchen Lin, Sanbao Su, Rachel Luo, Yuxiao Chen, Yan Wang, Marco Pavone, Fei Miao
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03100v3

Multimodal large language models (MLLMs) typically rely on a single late-layer feature from a frozen vision encoder, leaving the encoder's rich hierarchy of visual cues under-utilized. MLLMs still suffer from visually ungrounded hallucinations, often relying on language priors rather than image evidence. While many prior mitigation strategies operate on the text side, they leave the visual representation unchanged and do not exploit the rich hierarchy of features encoded across vision layers. Existing multi-layer fusion methods partially address this limitation but remain static, applying the same layer mixture regardless of the query. In this work, we introduce TGIF (Text-Guided Inter-layer Fusion), a lightweight module that treats encoder layers as depth-wise "experts" and predicts a prompt-dependent fusion of visual features. TGIF follows the principle of direct external fusion, requires no vision-encoder updates, and adds minimal overhead. Integrated into LLaVA-1.5-7B, TGIF provides consistent improvements across hallucination, OCR, and VQA benchmarks, while preserving or improving performance on ScienceQA, GQA, and MMBench. These results suggest that query-conditioned, hierarchy-aware fusion is an effective way to strengthen visual grounding and reduce hallucination in modern MLLMs. Code: https://github.com/Linchenchen/TGIF.

---

### Text-Guided Layer Fusion Mitigates Hallucination in Multimodal LLMs

Authors: Chenchen Lin, Sanbao Su, Rachel Luo, Yuxiao Chen, Yan Wang, Marco Pavone, Fei Miao
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03100v1

Multimodal large language models (MLLMs) typically rely on a single late-layer feature from a frozen vision encoder, leaving the encoder's rich hierarchy of visual cues under-utilized. MLLMs still suffer from visually ungrounded hallucinations, often relying on language priors rather than image evidence. While many prior mitigation strategies operate on the text side, they leave the visual representation unchanged and do not exploit the rich hierarchy of features encoded across vision layers. Existing multi-layer fusion methods partially address this limitation but remain static, applying the same layer mixture regardless of the query. In this work, we introduce TGIF (Text-Guided Inter-layer Fusion), a lightweight module that treats encoder layers as depth-wise "experts" and predicts a prompt-dependent fusion of visual features. TGIF follows the principle of direct external fusion, requires no vision-encoder updates, and adds minimal overhead. Integrated into LLaVA-1.5-7B, TGIF provides consistent improvements across hallucination, OCR, and VQA benchmarks, while preserving or improving performance on ScienceQA, GQA, and MMBench. These results suggest that query-conditioned, hierarchy-aware fusion is an effective way to strengthen visual grounding and reduce hallucination in modern MLLMs.

---

### Time-Aware Synthetic Control

Authors: Saeyoung Rho, Cyrus Illick, Samhitha Narasipura, Alberto Abadie, Daniel Hsu, Vishal Misra
Published: 2026-01-06
Categories: cs.LG, econ.EM, stat.ML
arXiv: https://arxiv.org/abs/2601.03099v1

The synthetic control (SC) framework is widely used for observational causal inference with time-series panel data. SC has been successful in diverse applications, but existing methods typically treat the ordering of pre-intervention time indices interchangeable. This invariance means they may not fully take advantage of temporal structure when strong trends are present. We propose Time-Aware Synthetic Control (TASC), which employs a state-space model with a constant trend while preserving a low-rank structure of the signal. TASC uses the Kalman filter and Rauch-Tung-Striebel smoother: it first fits a generative time-series model with expectation-maximization and then performs counterfactual inference. We evaluate TASC on both simulated and real-world datasets, including policy evaluation and sports prediction. Our results suggest that TASC offers advantages in settings with strong temporal trends and high levels of observation noise.

---

### From Muscle to Text with MyoText: sEMG to Text via Finger Classification and Transformer-Based Decoding

Authors: Meghna Roy Chowdhury, Shreyas Sen, Yi Ding
Published: 2026-01-06
Categories: cs.LG, cs.NE
arXiv: https://arxiv.org/abs/2601.03098v1

Surface electromyography (sEMG) provides a direct neural interface for decoding muscle activity and offers a promising foundation for keyboard-free text input in wearable and mixed-reality systems. Previous sEMG-to-text studies mainly focused on recognizing letters directly from sEMG signals, forming an important first step toward translating muscle activity into text. Building on this foundation, we present MyoText, a hierarchical framework that decodes sEMG signals to text through physiologically grounded intermediate stages. MyoText first classifies finger activations from multichannel sEMG using a CNN-BiLSTM-Attention model, applies ergonomic typing priors to infer letters, and reconstructs full sentences with a fine-tuned T5 transformer. This modular design mirrors the natural hierarchy of typing, linking muscle intent to language output and reducing the search space for decoding. Evaluated on 30 users from the emg2qwerty dataset, MyoText outperforms baselines by achieving 85.4% finger-classification accuracy, 5.4% character error rate (CER), and 6.5% word error rate (WER). Beyond accuracy gains, this methodology establishes a principled pathway from neuromuscular signals to text, providing a blueprint for virtual and augmented-reality typing interfaces that operate entirely without physical keyboards. By integrating ergonomic structure with transformer-based linguistic reasoning, MyoText advances the feasibility of seamless, wearable neural input for future ubiquitous computing environments.

---

### ATLAS: Verifier-Guided Adaptive Latent Activation Steering for Efficient LLM Reasoning

Authors: Tuc Nguyen, Thai Le
Published: 2026-01-06
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2601.03093v4

Recent work on activation and latent steering has demonstrated that modifying internal representations can effectively guide large language models (LLMs) toward improved reasoning and efficiency without updating model parameters. However, most existing approaches rely on fixed steering policies and static intervention strengths, which limit their robustness across problem instances and often result in over- or under-steering. We propose Adaptive Test-time Latent Steering (ATLAS), a lightweight framework that dynamically controls steering decisions at inference time using a trained, lightweight verifier over the latent states. Given intermediate hidden states, the verifier predicts the quality of ongoing reasoning and adaptively selects which steering action to apply, enabling per-example and per-step adjustment with minimal overhead. ATLAS provides a unified framework for combining learned latent verification with test-time activation steering, enabling adaptive reasoning control without additional LLM decoding or inference-time process reward model calls. Experiments on multiple mathematical and coding reasoning benchmarks show that ATLAS consistently outperforms both vanilla decoding and fixed steering baselines, achieving higher accuracy while substantially reducing test-time token usage. These results demonstrate that verifier-guided latent adaptation provides an effective and scalable mechanism for controlling reasoning efficiency without sacrificing solution quality. All source code will be publicly available.

---

### ATLAS: Adaptive Test-Time Latent Steering with External Verifiers for Enhancing LLMs Reasoning

Authors: Tuc Nguyen, Thai Le
Published: 2026-01-06
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2601.03093v1

Recent work on activation and latent steering has demonstrated that modifying internal representations can effectively guide large language models (LLMs) toward improved reasoning and efficiency without additional training. However, most existing approaches rely on fixed steering policies and static intervention strengths, which limit their robustness across problem instances and often result in over- or under-steering. We propose Adaptive Test-time Latent Steering, called (ATLAS), a task- specific framework that dynamically controls steering decisions at inference time using an external, lightweight latent verifier. Given intermediate hidden states, the verifier predicts the quality of ongoing reasoning and adaptively selects whether and how strongly to apply steering, enabling per-example and per-step adjustment with minimal overhead. To our knowledge, ATLAS is the first method to integrate learned latent verification into test-time steering for enhancing LLMs reasoning. Experiments on multiple mathematical reasoning benchmarks show that ATLAS consistently outperforms both vanilla decoding and fixed steering baselines, achieving higher accuracy while substantially reducing test-time token usage. These results demonstrate that verifier-guided latent adaptation provides an effective and scalable mechanism for controlling reasoning efficiency without sacrificing solution quality. All source code will be publicly available.

---

### LesionTABE: Equitable AI for Skin Lesion Detection

Authors: Rocio Mexia Diaz, Yasmin Greenway, Petru Manescu
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03090v1

Bias remains a major barrier to the clinical adoption of AI in dermatology, as diagnostic models underperform on darker skin tones. We present LesionTABE, a fairness-centric framework that couples adversarial debiasing with dermatology-specific foundation model embeddings. Evaluated across multiple datasets covering both malignant and inflammatory conditions, LesionTABE achieves over a 25\% improvement in fairness metrics compared to a ResNet-152 baseline, outperforming existing debiasing methods while simultaneously enhancing overall diagnostic accuracy. These results highlight the potential of foundation model debiasing as a step towards equitable clinical AI adoption.

---

### Faithfulness Evaluation for Decoder-only LLM Attributions with Controlled Retained Information

Authors: Xin Huang, Antoni B. Chan
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03089v2

Large Language Models (LLMs) are increasingly evaluated with input attribution methods, yet comparing such explanations remains challenging. Existing soft-perturbation faithfulness metrics, such as Soft-NC and Soft-NS, can conflate attribution quality with the number of words retained during perturbation: attribution methods with larger average scores may keep more words and therefore obtain inflated scores. To address this issue, we propose $π$-Soft-NC and $π$-Soft-NS, an evaluation framework that compares attribution methods under the same expected retaining probability, thus controlling the number of retained words. We further introduce Grad-ELLM, a gradient-based attribution method tailored to autoregressive decoder-only LLMs, which combines gradient-derived channel importance with attention-derived token importance at each decoding step. Experiments on classification and open-generation tasks with Llama and Mistral show that Grad-ELLM achieves strong comprehensiveness-oriented faithfulness under $π$-Soft-NC, while there is no dominant method under $π$-Soft-NS. Our evaluation metric serves as a rigorous framework to compare XAI methods for LLMs, which will support progress in the field.

---

### Grad-ELLM: Gradient-based Explanations for Decoder-only LLMs

Authors: Xin Huang, Antoni B. Chan
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03089v1

Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks, yet their black-box nature raises concerns about transparency and faithfulness. Input attribution methods aim to highlight each input token's contributions to the model's output, but existing approaches are typically model-agnostic, and do not focus on transformer-specific architectures, leading to limited faithfulness. To address this, we propose Grad-ELLM, a gradient-based attribution method for decoder-only transformer-based LLMs. By aggregating channel importance from gradients of the output logit with respect to attention layers and spatial importance from attention maps, Grad-ELLM generates heatmaps at each generation step without requiring architectural modifications. Additionally, we introduce two faithfulneses metrics $π$-Soft-NC and $π$-Soft-NS, which are modifications of Soft-NC/NS that provide fairer comparisons by controlling the amount of information kept when perturbing the text. We evaluate Grad-ELLM on sentiment classification, question answering, and open-generation tasks using different models. Experiment results show that Grad-ELLM consistently achieves superior faithfulness than other attribution methods.

---

### Audit Me If You Can: Query-Efficient Active Fairness Auditing of Black-Box LLMs

Authors: David Hartmann, Lena Pohlmann, Lelia Hanslik, Noah Gießing, Bettina Berendt, Pieter Delobelle
Published: 2026-01-06
Categories: cs.LG, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2601.03087v2

Large Language Models (LLMs) exhibit systematic biases across demographic groups. Auditing is proposed as an accountability tool for black-box LLM applications, but suffers from resource-intensive query access. We conceptualise auditing as uncertainty estimation over a target fairness metric and introduce BAFA, the Bounded Active Fairness Auditor for query-efficient auditing of black-box LLMs. BAFA maintains a version space of surrogate models consistent with queried scores and computes uncertainty intervals for fairness metrics (e.g., $Δ$ AUC) via constrained empirical risk minimisation. Active query selection narrows these intervals to reduce estimation error. We evaluate BAFA on two standard fairness dataset case studies: \textsc{CivilComments} and \textsc{Bias-in-Bios}, comparing against stratified sampling, power sampling, and ablations. BAFA achieves target error thresholds with up to 40$\times$ fewer queries than stratified sampling (e.g., 144 vs 5,956 queries at $\varepsilon=0.02$ for \textsc{CivilComments}) for tight thresholds, demonstrates substantially better performance over time, and shows lower variance across runs. These results suggest that active sampling can reduce resources needed for independent fairness auditing with LLMs, supporting continuous model evaluations.

---

### Audit Me If You Can: Query-Efficient Active Fairness Auditing of Black-Box LLMs

Authors: David Hartmann, Lena Pohlmann, Lelia Hanslik, Noah Gießing, Bettina Berendt, Pieter Delobelle
Published: 2026-01-06
Categories: cs.LG, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2601.03087v1

Large Language Models (LLMs) exhibit systematic biases across demographic groups. Auditing is proposed as an accountability tool for black-box LLM applications, but suffers from resource-intensive query access. We conceptualise auditing as uncertainty estimation over a target fairness metric and introduce BAFA, the Bounded Active Fairness Auditor for query-efficient auditing of black-box LLMs. BAFA maintains a version space of surrogate models consistent with queried scores and computes uncertainty intervals for fairness metrics (e.g., $Δ$ AUC) via constrained empirical risk minimisation. Active query selection narrows these intervals to reduce estimation error. We evaluate BAFA on two standard fairness dataset case studies: \textsc{CivilComments} and \textsc{Bias-in-Bios}, comparing against stratified sampling, power sampling, and ablations. BAFA achieves target error thresholds with up to 40$\times$ fewer queries than stratified sampling (e.g., 144 vs 5,956 queries at $\varepsilon=0.02$ for \textsc{CivilComments}) for tight thresholds, demonstrates substantially better performance over time, and shows lower variance across runs. These results suggest that active sampling can reduce resources needed for independent fairness auditing with LLMs, supporting continuous model evaluations.

---

### Real-Time Adaptive Anomaly Detection in Industrial IoT Environments

Authors: Mahsa Raeiszadeh, Amin Ebrahimzadeh, Roch H. Glitho, Johan Eker, Raquel A. F. Mini
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03085v1

To ensure reliability and service availability, next-generation networks are expected to rely on automated anomaly detection systems powered by advanced machine learning methods with the capability of handling multi-dimensional data. Such multi-dimensional, heterogeneous data occurs mostly in today's industrial Internet of Things (IIoT), where real-time detection of anomalies is critical to prevent impending failures and resolve them in a timely manner. However, existing anomaly detection methods often fall short of effectively coping with the complexity and dynamism of multi-dimensional data streams in IIoT. In this paper, we propose an adaptive method for detecting anomalies in IIoT streaming data utilizing a multi-source prediction model and concept drift adaptation. The proposed anomaly detection algorithm merges a prediction model into a novel drift adaptation method resulting in accurate and efficient anomaly detection that exhibits improved scalability. Our trace-driven evaluations indicate that the proposed method outperforms the state-of-the-art anomaly detection methods by achieving up to an 89.71% accuracy (in terms of Area under the Curve (AUC)) while meeting the given efficiency and scalability requirements.

---

### Learning to Diagnose and Correct Moral Errors: Beyond Shallow Heuristics in Moral Alignment

Authors: Bocheng Chen, Xi Chen, Han Zi, Haitao Mao, Zimo Qi, Xitong Zhang, Kristen Johnson, Guangliang Liu
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03079v6

Existing approaches to moral value alignment are primarily set out to align LLMs' generation with the distributions of morally appropriate language, which has seen good progress. However, these approaches are often brittle, heavily rely on shallow heuristics, and reduce performance in out-of-the-distribution tasks. In other words, the learning paradigm underlying existing approaches teaches LLMs what morally (in)appropriate language looks like, but not why it is morally (in)appropriate. In this paper, we address this challenge by developing pragmatic inference-driven methods to facilitate LLMs' learning of how to diagnose and correct moral errors, thereby enabling them to generate morally appropriate language. Pragmatic inference is the reasoning process of deriving (implied) meanings -- a famous concept in linguistics. Our methods vary the inference procedures by the inferential load of different moral discourses, rather than modelling their diverse and complex semantic distributions separately. Empirical results demonstrate that our approach improves moral value alignment in LLMs and generalizes effectively across tasks.

---

### Learning to Diagnose and Correct Moral Errors: Towards Enhancing Moral Sensitivity in Large Language Models

Authors: Bocheng Chen, Han Zi, Xi Chen, Xitong Zhang, Kristen Johnson, Guangliang Liu
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03079v1

Moral sensitivity is fundamental to human moral competence, as it guides individuals in regulating everyday behavior. Although many approaches seek to align large language models (LLMs) with human moral values, how to enable them morally sensitive has been extremely challenging. In this paper, we take a step toward answering the question: how can we enhance moral sensitivity in LLMs? Specifically, we propose two pragmatic inference methods that faciliate LLMs to diagnose morally benign and hazardous input and correct moral errors, whereby enhancing LLMs' moral sensitivity. A central strength of our pragmatic inference methods is their unified perspective: instead of modeling moral discourses across semantically diverse and complex surface forms, they offer a principled perspective for designing pragmatic inference procedures grounded in their inferential loads. Empirical evidence demonstrates that our pragmatic methods can enhance moral sensitivity in LLMs and achieves strong performance on representative morality-relevant benchmarks.

---

### Understanding Multi-Agent Reasoning with Large Language Models for Cartoon VQA

Authors: Tong Wu, Thanet Markchom
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03073v1

Visual Question Answering (VQA) for stylised cartoon imagery presents challenges, such as interpreting exaggerated visual abstraction and narrative-driven context, which are not adequately addressed by standard large language models (LLMs) trained on natural images. To investigate this issue, a multi-agent LLM framework is introduced, specifically designed for VQA tasks in cartoon imagery. The proposed architecture consists of three specialised agents: visual agent, language agent and critic agent, which work collaboratively to support structured reasoning by integrating visual cues and narrative context. The framework was systematically evaluated on two cartoon-based VQA datasets: Pororo and Simpsons. Experimental results provide a detailed analysis of how each agent contributes to the final prediction, offering a deeper understanding of LLM-based multi-agent behaviour in cartoon VQA and multimodal inference.

---

### Joint Encoding of KV-Cache Blocks for Scalable LLM Serving

Authors: Joseph Kampeas, Emir Haleva
Published: 2026-01-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.03067v1

Modern large language models (LLMs) drive interactive AI systems but are bottlenecked by the memory-heavy growth of key-value (KV) caches, which limits real-time throughput under concurrent loads. Existing KV-cache compression methods rely on rigid heuristics, disrupt tensor layouts, or require specialized compute, hindering scalability and deployment. We propose joint encoding of KV-cache blocks, which fuses similar blocks across requests and input chunks into shared representations while preserving standard cache structure. This alleviates the KV-cache memory bottleneck, supporting high-concurrency serving without specialized hardware. Theoretically, we analyze the rate-distortion tradeoff of fused cache blocks under a Poisson process model. Empirically, our method achieves up to 4.38 $\times$ KV-cache compression with negligible accuracy loss across diverse LLMs and benchmarks, outperforming recent structured and adaptive compression baselines. In real LLM serving, joint encoding improves the token throughput by $\sim$40\% on a single-machine vLLM benchmark, demonstrating substantial gains in inference throughput. Code is available at https://github.com/sef1/kv_fast_fusion kv_joint_encoding.

---

### Do LLMs Encode Functional Importance of Reasoning Tokens?

Authors: Janvijay Singh, Dilek Hakkani-Tür
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03066v3

Large language models solve complex tasks by generating long reasoning chains, achieving higher accuracy at the cost of increased computational cost and reduced ability to isolate functionally relevant reasoning. Prior work on compact reasoning shortens such chains through probabilistic sampling, heuristics, or supervision from frontier models, but offers limited insight into whether models internally encode token-level functional importance for answer generation. We address this gap diagnostically and propose greedy pruning, a likelihood-preserving deletion procedure that iteratively removes reasoning tokens whose removal minimally degrades model likelihood under a specified objective, yielding length-controlled reasoning chains. We evaluate pruned reasoning in a distillation framework and show that students trained on pruned chains outperform a frontier-model-supervised compression baseline at matched reasoning lengths. Finally, our analysis reveals systematic pruning patterns and shows that attention scores can predict greedy pruning ranks, further suggesting that models encode a nontrivial functional importance structure over reasoning tokens.

---

### Do LLMs Encode Functional Importance of Reasoning Tokens?

Authors: Janvijay Singh, Dilek Hakkani-Tür
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03066v1

Large language models solve complex tasks by generating long reasoning chains, achieving higher accuracy at the cost of increased computational cost and reduced ability to isolate functionally relevant reasoning. Prior work on compact reasoning shortens such chains through probabilistic sampling, heuristics, or supervision from frontier models, but offers limited insight into whether models internally encode token-level functional importance for answer generation. We address this gap diagnostically and propose greedy pruning, a likelihood-preserving deletion procedure that iteratively removes reasoning tokens whose removal minimally degrades model likelihood under a specified objective, yielding length-controlled reasoning chains. We evaluate pruned reasoning in a distillation framework and show that students trained on pruned chains outperform a frontier-model-supervised compression baseline at matched reasoning lengths. Finally, our analysis reveals systematic pruning patterns and shows that attention scores can predict greedy pruning ranks, further suggesting that models encode a nontrivial functional importance structure over reasoning tokens.

---

### Explainable Fuzzy GNNs for Leak Detection in Water Distribution Networks

Authors: Qusai Khaled, Pasquale De Marinis, Moez Louati, David Ferras, Laura Genga, Uzay Kaymak
Published: 2026-01-06
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03062v1

Timely leak detection in water distribution networks is critical for conserving resources and maintaining operational efficiency. Although Graph Neural Networks (GNNs) excel at capturing spatial-temporal dependencies in sensor data, their black-box nature and the limited work on graph-based explainable models for water networks hinder practical adoption. We propose an explainable GNN framework that integrates mutual information to identify critical network regions and fuzzy logic to provide clear, rule-based explanations for node classification tasks. After benchmarking several GNN architectures, we selected the generalized graph convolution network (GENConv) for its superior performance and developed a fuzzy-enhanced variant that offers intuitive explanations for classified leak locations. Our fuzzy graph neural network (FGENConv) achieved Graph F1 scores of 0.889 for detection and 0.814 for localization, slightly below the crisp GENConv 0.938 and 0.858, respectively. Yet it compensates by providing spatially localized, fuzzy rule-based explanations. By striking the right balance between precision and explainability, the proposed fuzzy network could enable hydraulic engineers to validate predicted leak locations, conserve human resources, and optimize maintenance strategies. The code is available at github.com/pasqualedem/GNNLeakDetection.

---

### Fine-Grained Generalization via Structuralizing Concept and Feature Space into Commonality, Specificity and Confounding

Authors: Zhen Wang, Jiaojiao Zhao, Qilong Wang, Yongfeng Dong, Wenlong Yu
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03056v1

Fine-Grained Domain Generalization (FGDG) presents greater challenges than conventional domain generalization due to the subtle inter-class differences and relatively pronounced intra-class variations inherent in fine-grained recognition tasks. Under domain shifts, the model becomes overly sensitive to fine-grained cues, leading to the suppression of critical features and a significant drop in performance. Cognitive studies suggest that humans classify objects by leveraging both common and specific attributes, enabling accurate differentiation between fine-grained categories. However, current deep learning models have yet to incorporate this mechanism effectively. Inspired by this mechanism, we propose Concept-Feature Structuralized Generalization (CFSG). This model explicitly disentangles both the concept and feature spaces into three structured components: common, specific, and confounding segments. To mitigate the adverse effects of varying degrees of distribution shift, we introduce an adaptive mechanism that dynamically adjusts the proportions of common, specific, and confounding components. In the final prediction, explicit weights are assigned to each pair of components. Extensive experiments on three single-source benchmark datasets demonstrate that CFSG achieves an average performance improvement of 9.87% over baseline models and outperforms existing state-of-the-art methods by an average of 3.08%. Additionally, explainability analysis validates that CFSG effectively integrates multi-granularity structured knowledge and confirms that feature structuralization facilitates the emergence of concept structuralization.

---

### IBISAgent: Reinforcing Pixel-Level Visual Reasoning in MLLMs for Universal Biomedical Object Referring and Segmentation

Authors: Yankai Jiang, Qiaoru Li, Binlu Xu, Haoran Sun, Chao Ding, Junting Dong, Yuxiang Cai, Xuhong Zhang, Jianwei Yin
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03054v4

Recent research on medical MLLMs has gradually shifted its focus from image-level understanding to fine-grained, pixel-level comprehension. Although segmentation serves as the foundation for pixel-level understanding, existing approaches face two major challenges. First, they introduce implicit segmentation tokens and require simultaneous fine-tuning of both the MLLM and external pixel decoders, which increases the risk of catastrophic forgetting and limits generalization to out-of-domain scenarios. Second, most methods rely on single-pass reasoning and lack the capability to iteratively refine segmentation results, leading to suboptimal performance. To overcome these limitations, we propose a novel agentic MLLM, named IBISAgent, that reformulates segmentation as a vision-centric, multi-step decision-making process. IBISAgent enables MLLMs to generate interleaved reasoning and text-based click actions, invoke segmentation tools, and produce high-quality masks without architectural modifications. By iteratively performing multi-step visual reasoning on masked image features, IBISAgent naturally supports mask refinement and promotes the development of pixel-level visual reasoning capabilities. We further design a two-stage training framework consisting of cold-start supervised fine-tuning and agentic reinforcement learning with tailored, fine-grained rewards, enhancing the model's robustness in complex medical referring and reasoning segmentation tasks. Extensive experiments demonstrate that IBISAgent consistently outperforms both closed-source and open-source SOTA methods.

---

### IBISAgent: Reinforcing Pixel-Level Visual Reasoning in MLLMs for Universal Biomedical Object Referring and Segmentation

Authors: Yankai Jiang, Qiaoru Li, Binlu Xu, Haoran Sun, Chao Ding, Junting Dong, Yuxiang Cai, Xuhong Zhang, Jianwei Yin
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03054v1

Recent research on medical MLLMs has gradually shifted its focus from image-level understanding to fine-grained, pixel-level comprehension. Although segmentation serves as the foundation for pixel-level understanding, existing approaches face two major challenges. First, they introduce implicit segmentation tokens and require simultaneous fine-tuning of both the MLLM and external pixel decoders, which increases the risk of catastrophic forgetting and limits generalization to out-of-domain scenarios. Second, most methods rely on single-pass reasoning and lack the capability to iteratively refine segmentation results, leading to suboptimal performance. To overcome these limitations, we propose a novel agentic MLLM, named IBISAgent, that reformulates segmentation as a vision-centric, multi-step decision-making process. IBISAgent enables MLLMs to generate interleaved reasoning and text-based click actions, invoke segmentation tools, and produce high-quality masks without architectural modifications. By iteratively performing multi-step visual reasoning on masked image features, IBISAgent naturally supports mask refinement and promotes the development of pixel-level visual reasoning capabilities. We further design a two-stage training framework consisting of cold-start supervised fine-tuning and agentic reinforcement learning with tailored, fine-grained rewards, enhancing the model's robustness in complex medical referring and reasoning segmentation tasks. Extensive experiments demonstrate that IBISAgent consistently outperforms both closed-source and open-source SOTA methods. All datasets, code, and trained models will be released publicly.

---

### Detecting Hallucinations in Retrieval-Augmented Generation via Semantic-level Internal Reasoning Graph

Authors: Jianpeng Hu, Yanzeng Li, Jialun Zhong, Wenfa Qi, Lei Zou
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03052v1

The Retrieval-augmented generation (RAG) system based on Large language model (LLM) has made significant progress. It can effectively reduce factuality hallucinations, but faithfulness hallucinations still exist. Previous methods for detecting faithfulness hallucinations either neglect to capture the models' internal reasoning processes or handle those features coarsely, making it difficult for discriminators to learn. This paper proposes a semantic-level internal reasoning graph-based method for detecting faithfulness hallucination. Specifically, we first extend the layer-wise relevance propagation algorithm from the token level to the semantic level, constructing an internal reasoning graph based on attribution vectors. This provides a more faithful semantic-level representation of dependency. Furthermore, we design a general framework based on a small pre-trained language model to utilize the dependencies in LLM's reasoning for training and hallucination detection, which can dynamically adjust the pass rate of correct samples through a threshold. Experimental results demonstrate that our method achieves better overall performance compared to state-of-the-art baselines on RAGTruth and Dolly-15k.

---

### Temporal Graph Network: Hallucination Detection in Multi-Turn Conversation

Authors: Vidhi Rathore, Sambu Aneesh, Himanshu Singh
Published: 2026-01-06
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2601.03051v1

Hallucinations can be produced by conversational AI systems, particularly in multi-turn conversations where context changes and contradictions may eventually surface. By representing the entire conversation as a temporal graph, we present a novel graph-based method for detecting dialogue-level hallucinations. Our framework models each dialogue as a node, encoding it using a sentence transformer. We explore two different ways of connectivity: i) shared-entity edges, which connect turns that refer to the same entities; ii) temporal edges, which connect contiguous turns in the conversation. Message-passing is used to update the node embeddings, allowing flow of information between related nodes. The context-aware node embeddings are then combined using attention pooling into a single vector, which is then passed on to a classifier to determine the presence and type of hallucinations. We demonstrate that our method offers slightly improved performance over existing methods. Further, we show the attention mechanism can be used to justify the decision making process. The code and model weights are made available at: https://github.com/sambuaneesh/anlp-project.

---

### On the Intrinsic Limits of Transformer Image Embeddings in Non-Solvable Spatial Reasoning

Authors: Siyi Lyu, Quan Liu, Feng Yan
Published: 2026-01-06
Categories: cs.CV, cs.AI, cs.CC
arXiv: https://arxiv.org/abs/2601.03048v2

Vision Transformers (ViTs) excel in semantic recognition but exhibit systematic failures in spatial reasoning tasks such as mental rotation. While often attributed to data scale, this work argues that the limitation arises from the intrinsic circuit complexity of the architecture. By formalizing spatial understanding as learning a Group Homomorphism Problem -- where latent embeddings preserve the algebraic structure of physical transformations acting on images -- we identify a fundamental computational bottleneck. Specifically, for non-solvable groups (e.g., $\mathrm{SO}(3)$), maintaining such structure-preserving embeddings is lowerbounded by the Word Problem, which is $\mathsf{NC^1}$-complete. In contrast, constant-depth ViTs with polynomial precision are strictly bounded by the complexity class $\mathsf{TC^0}$. Under the standard conjecture $\mathsf{TC^0} \subsetneq \mathsf{NC^1}$, a complexity boundary emerges: constant-depth architectures lack the logical depth required to capture non-solvable spatial structures in a single forward pass. To empirically validate this theoretical gap, we propose the Latent Space Algebra (LSA) benchmark, which reveals a significant degradation in ViT representations as the compositional depth of non-solvable tasks increases.

---

### On the Intrinsic Limits of Transformer Image Embeddings in Non-Solvable Spatial Reasoning

Authors: Siyi Lyu, Quan Liu, Feng Yan
Published: 2026-01-06
Categories: cs.CV, cs.AI, cs.CC
arXiv: https://arxiv.org/abs/2601.03048v1

Vision Transformers (ViTs) excel in semantic recognition but exhibit systematic failures in spatial reasoning tasks such as mental rotation. While often attributed to data scale, we propose that this limitation arises from the intrinsic circuit complexity of the architecture. We formalize spatial understanding as learning a Group Homomorphism: mapping image sequences to a latent space that preserves the algebraic structure of the underlying transformation group. We demonstrate that for non-solvable groups (e.g., the 3D rotation group $\mathrm{SO}(3)$), maintaining such a structure-preserving embedding is computationally lower-bounded by the Word Problem, which is $\mathsf{NC^1}$-complete. In contrast, we prove that constant-depth ViTs with polynomial precision are strictly bounded by $\mathsf{TC^0}$. Under the conjecture $\mathsf{TC^0} \subsetneq \mathsf{NC^1}$, we establish a complexity boundary: constant-depth ViTs fundamentally lack the logical depth to efficiently capture non-solvable spatial structures. We validate this complexity gap via latent-space probing, demonstrating that ViT representations suffer a structural collapse on non-solvable tasks as compositional depth increases.

---

### When the Coffee Feature Activates on Coffins: An Analysis of Feature Extraction and Steering for Mechanistic Interpretability

Authors: Raphael Ronge, Markus Maier, Frederick Eberhardt
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03047v1

Recent work by Anthropic on Mechanistic interpretability claims to understand and control Large Language Models by extracting human-interpretable features from their neural activation patterns using sparse autoencoders (SAEs). If successful, this approach offers one of the most promising routes for human oversight in AI safety. We conduct an initial stress-test of these claims by replicating their main results with open-source SAEs for Llama 3.1. While we successfully reproduce basic feature extraction and steering capabilities, our investigation suggests that major caution is warranted regarding the generalizability of these claims. We find that feature steering exhibits substantial fragility, with sensitivity to layer selection, steering magnitude, and context. We observe non-standard activation behavior and demonstrate the difficulty to distinguish thematically similar features from one another. While SAE-based interpretability produces compelling demonstrations in selected cases, current methods often fall short of the systematic reliability required for safety-critical applications. This suggests a necessary shift in focus from prioritizing interpretability of internal representations toward reliable prediction and control of model output. Our work contributes to a more nuanced understanding of what mechanistic interpretability has achieved and highlights fundamental challenges for AI safety that remain unresolved.

---

### Motion Blur Robust Wheat Pest Damage Detection with Dynamic Fuzzy Feature Fusion

Authors: Han Zhang, Yanwei Wang, Fang Li, Hongjun Wang
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03046v1

Motion blur caused by camera shake produces ghosting artifacts that substantially degrade edge side object detection. Existing approaches either suppress blur as noise and lose discriminative structure, or apply full image restoration that increases latency and limits deployment on resource constrained devices. We propose DFRCP, a Dynamic Fuzzy Robust Convolutional Pyramid, as a plug in upgrade to YOLOv11 for blur robust detection. DFRCP enhances the YOLOv11 feature pyramid by combining large scale and medium scale features while preserving native representations, and by introducing Dynamic Robust Switch units that adaptively inject fuzzy features to strengthen global perception under jitter. Fuzzy features are synthesized by rotating and nonlinearly interpolating multiscale features, then merged through a transparency convolution that learns a content adaptive trade off between original and fuzzy cues. We further develop a CUDA parallel rotation and interpolation kernel that avoids boundary overflow and delivers more than 400 times speedup, making the design practical for edge deployment. We train with paired supervision on a private wheat pest damage dataset of about 3,500 images, augmented threefold using two blur regimes, uniform image wide motion blur and bounding box confined rotational blur. On blurred test sets, YOLOv11 with DFRCP achieves about 10.4 percent higher accuracy than the YOLOv11 baseline with only a modest training time overhead, reducing the need for manual filtering after data collection.

---

### Lil: Less is Less When Applying Post-Training Sparse-Attention Algorithms in Long-Decode Stage

Authors: Junhao Hu, Fangze Li, Mingtao Xu, Feifan Meng, Shiju Zhao, Tiancheng Hu, Ting Peng, Anmin Liu, Wenrui Huang, Chenxu Liu, Ziyue Hua, Tao Xie
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03043v4

Large language models (LLMs) demonstrate strong capabilities across a wide range of complex tasks and are increasingly deployed at scale, placing significant demands on inference efficiency. Prior work typically decomposes inference into prefill and decode stages, with the decode stage dominating total latency. To reduce time and memory complexity in the decode stage, a line of work introduces sparse-attention algorithms. In this paper, we show, both empirically and theoretically, that sparse attention can paradoxically increase end-to-end complexity: information loss often induces significantly longer sequences, a phenomenon we term ``Less is Less'' (Lil). To mitigate the Lil problem, we propose an early-stopping algorithm that detects the threshold where information loss exceeds information gain during sparse decoding. Our early-stopping algorithm reduces token consumption by up to 90% with a marginal accuracy degradation of less than 2% across reasoning-intensive benchmarks.

---

### Lil: Less is Less When Applying Post-Training Sparse-Attention Algorithms in Long-Decode Stage

Authors: Junhao Hu, Fangze Li, Mingtao Xu, Feifan Meng, Shiju Zhao, Tiancheng Hu, Ting Peng, Anmin Liu, Wenrui Huang, Chenxu Liu, Ziyue Hua, Tao Xie
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03043v1

Large language models (LLMs) demonstrate strong capabilities across a wide range of complex tasks and are increasingly deployed at scale, placing significant demands on inference efficiency. Prior work typically decomposes inference into prefill and decode stages, with the decode stage dominating total latency. To reduce time and memory complexity in the decode stage, a line of work introduces sparse-attention algorithms. In this paper, we show, both empirically and theoretically, that sparse attention can paradoxically increase end-to-end complexity: information loss often induces significantly longer sequences, a phenomenon we term ``Less is Less'' (Lil). To mitigate the Lil problem, we propose an early-stopping algorithm that detects the threshold where information loss exceeds information gain during sparse decoding. Our early-stopping algorithm reduces token consumption by up to 90% with a marginal accuracy degradation of less than 2% across reasoning-intensive benchmarks.

---

### BaseCal: Unsupervised Confidence Calibration via Base Model Signals

Authors: Hexiang Tan, Wanli Yang, Junwei Zhang, Xin Chen, Rui Tang, Du Su, Jingang Wang, Yuanzhuo Wang, Fei Sun, Xueqi Cheng
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03042v4

Reliable confidence is essential for trusting the outputs of LLMs, yet widely deployed post-trained LLMs (PoLLMs) typically compromise this trust with severe overconfidence. In contrast, we observe that their corresponding base LLMs often remain well-calibrated. This naturally motivates us to calibrate PoLLM confidence using the base LLM as a reference. This work proposes two ways to achieve this. A straightforward solution, BaseCal-ReEval, evaluates PoLLM's responses by feeding them into the base LLM to get average probabilities as confidence. While effective, this approach introduces additional inference overhead. To address this, we propose BaseCal-Proj, which trains a lightweight projection to map the final-layer hidden states of PoLLMs back to those of their base LLMs. These projected states are then processed by the base LLM's output layer to derive base-calibrated confidence for PoLLM's responses. Notably, BaseCal is an unsupervised, plug-and-play solution that operates without human labels or LLM modifications. Experiments across five datasets and three LLM families demonstrate the effectiveness of BaseCal, reducing Expected Calibration Error (ECE) by an average of 42.90\% compared to the best unsupervised baselines.

---

### BaseCal: Unsupervised Confidence Calibration via Base Model Signals

Authors: Hexiang Tan, Wanli Yang, Junwei Zhang, Xin Chen, Rui Tang, Du Su, Jingang Wang, Yuanzhuo Wang, Fei Sun, Xueqi Cheng
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03042v1

Reliable confidence is essential for trusting the outputs of LLMs, yet widely deployed post-trained LLMs (PoLLMs) typically compromise this trust with severe overconfidence. In contrast, we observe that their corresponding base LLMs often remain well-calibrated. This naturally motivates us to calibrate PoLLM confidence using the base LLM as a reference. This work proposes two ways to achieve this. A straightforward solution, BaseCal-ReEval, evaluates PoLLM's responses by feeding them into the base LLM to get average probabilities as confidence. While effective, this approach introduces additional inference overhead. To address this, we propose BaseCal-Proj, which trains a lightweight projection to map the final-layer hidden states of PoLLMs back to those of their base LLMs. These projected states are then processed by the base LLM's output layer to derive base-calibrated confidence for PoLLM's responses. Notably, BaseCal is an unsupervised, plug-and-play solution that operates without human labels or LLM modifications. Experiments across five datasets and three LLM families demonstrate the effectiveness of BaseCal, reducing Expected Calibration Error (ECE) by an average of 42.90\% compared to the best unsupervised baselines.

---

### PiDR: Physics-Informed Inertial Dead Reckoning for Autonomous Platforms

Authors: Arup Kumar Sahoo, Itzik Klein
Published: 2026-01-06
Categories: cs.RO, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03040v2

A fundamental requirement for full autonomy is the ability to sustain accurate navigation in the absence of external data, such as GNSS signals or visual information. In these challenging environments, the platform must rely exclusively on inertial sensors, leading to pure inertial navigation. However, the inherent noise and other error terms of the inertial sensors in such real-world scenarios will cause the navigation solution to drift over time. Although conventional deep-learning models have emerged as a possible approach to inertial navigation, they are inherently black-box in nature. Furthermore, they struggle to learn effectively with limited supervised sensor data and often fail to preserve physical principles. To address these limitations, we propose PiDR, a physics-informed inertial dead-reckoning framework for autonomous platforms in situations of pure inertial navigation. PiDR offers transparency by explicitly integrating inertial navigation principles into the network training process through the physics-informed residual component. PiDR plays a crucial role in mitigating abrupt trajectory deviations even under limited or sparse supervision. We evaluated PiDR on real-world datasets collected by a mobile robot and an autonomous underwater vehicle. We obtained more than 29% positioning improvement in both datasets, demonstrating the ability of PiDR to generalize different platforms operating in various environments and dynamics. Thus, PiDR offers a robust, lightweight, yet effective architecture and can be deployed on resource-constrained platforms, enabling real-time pure inertial navigation in adverse scenarios.

---

### PiDR: Physics-Informed Inertial Dead Reckoning for Autonomous Platforms

Authors: Arup Kumar Sahoo, Itzik Klein
Published: 2026-01-06
Categories: cs.RO, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03040v1

A fundamental requirement for full autonomy is the ability to sustain accurate navigation in the absence of external data, such as GNSS signals or visual information. In these challenging environments, the platform must rely exclusively on inertial sensors, leading to pure inertial navigation. However, the inherent noise and other error terms of the inertial sensors in such real-world scenarios will cause the navigation solution to drift over time. Although conventional deep-learning models have emerged as a possible approach to inertial navigation, they are inherently black-box in nature. Furthermore, they struggle to learn effectively with limited supervised sensor data and often fail to preserve physical principles. To address these limitations, we propose PiDR, a physics-informed inertial dead-reckoning framework for autonomous platforms in situations of pure inertial navigation. PiDR offers transparency by explicitly integrating inertial navigation principles into the network training process through the physics-informed residual component. PiDR plays a crucial role in mitigating abrupt trajectory deviations even under limited or sparse supervision. We evaluated PiDR on real-world datasets collected by a mobile robot and an autonomous underwater vehicle. We obtained more than 29% positioning improvement in both datasets, demonstrating the ability of PiDR to generalize different platforms operating in various environments and dynamics. Thus, PiDR offers a robust, lightweight, yet effective architecture and can be deployed on resource-constrained platforms, enabling real-time pure inertial navigation in adverse scenarios.

---

### Validating Generalist Robots with Situation Calculus and STL Falsification

Authors: Changwen Li, Rongjie Yan, Chih-Hong Cheng, Jian Zhang
Published: 2026-01-06
Categories: cs.RO, cs.AI
arXiv: https://arxiv.org/abs/2601.03038v1

Generalist robots are becoming a reality, capable of interpreting natural language instructions and executing diverse operations. However, their validation remains challenging because each task induces its own operational context and correctness specification, exceeding the assumptions of traditional validation methods. We propose a two-layer validation framework that combines abstract reasoning with concrete system falsification. At the abstract layer, situation calculus models the world and derives weakest preconditions, enabling constraint-aware combinatorial testing to systematically generate diverse, semantically valid world-task configurations with controllable coverage strength. At the concrete layer, these configurations are instantiated for simulation-based falsification with STL monitoring. Experiments on tabletop manipulation tasks show that our framework effectively uncovers failure cases in the NVIDIA GR00T controller, demonstrating its promise for validating general-purpose robot autonomy.

---

### NorwAI's Large Language Models: Technical Report

Authors: Jon Atle Gulla, Peng Liu, Lemei Zhang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03034v2

Norwegian, spoken by approximately five million people, remains underrepresented in many of the most significant breakthroughs in Natural Language Processing (NLP). To address this gap, the NorLLM team at NorwAI has developed a family of models specifically tailored to Norwegian and other Scandinavian languages, building on diverse Transformer-based architectures such as GPT, Mistral, Llama2, Mixtral and Magistral. These models are either pretrained from scratch or continually pretrained on 25B - 88.45B tokens, using a Norwegian-extended tokenizer and advanced post-training strategies to optimize performance, enhance robustness, and improve adaptability across various real-world tasks. Notably, instruction-tuned variants (e.g., Mistral-7B-Instruct and Mixtral-8x7B-Instruct) showcase strong assistant-style capabilities, underscoring their potential for practical deployment in interactive and domain-specific applications. The NorwAI large language models are openly available to Nordic organizations, companies and students for both research and experimental use. This report provides detailed documentation of the model architectures, training data, tokenizer design, fine-tuning strategies, deployment, and evaluations.

---

### NorwAI's Large Language Models: Technical Report

Authors: Jon Atle Gulla, Peng Liu, Lemei Zhang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03034v1

Norwegian, spoken by approximately five million people, remains underrepresented in many of the most significant breakthroughs in Natural Language Processing (NLP). To address this gap, the NorLLM team at NorwAI has developed a family of models specifically tailored to Norwegian and other Scandinavian languages, building on diverse Transformer-based architectures such as GPT, Mistral, Llama2, Mixtral and Magistral. These models are either pretrained from scratch or continually pretrained on 25B - 88.45B tokens, using a Norwegian-extended tokenizer and advanced post-training strategies to optimize performance, enhance robustness, and improve adaptability across various real-world tasks. Notably, instruction-tuned variants (e.g., Mistral-7B-Instruct and Mixtral-8x7B-Instruct) showcase strong assistant-style capabilities, underscoring their potential for practical deployment in interactive and domain-specific applications. The NorwAI large language models are openly available to Nordic organizations, companies and students for both research and experimental use. This report provides detailed documentation of the model architectures, training data, tokenizer design, fine-tuning strategies, deployment, and evaluations.

---

### Causal Manifold Fairness: Enforcing Geometric Invariance in Representation Learning

Authors: Vidhi Rathore
Published: 2026-01-06
Categories: cs.LG, cs.AI, cs.CY
arXiv: https://arxiv.org/abs/2601.03032v1

Fairness in machine learning is increasingly critical, yet standard approaches often treat data as static points in a high-dimensional space, ignoring the underlying generative structure. We posit that sensitive attributes (e.g., race, gender) do not merely shift data distributions but causally warp the geometry of the data manifold itself. To address this, we introduce Causal Manifold Fairness (CMF), a novel framework that bridges causal inference and geometric deep learning. CMF learns a latent representation where the local Riemannian geometry, defined by the metric tensor and curvature, remains invariant under counterfactual interventions on sensitive attributes. By enforcing constraints on the Jacobian and Hessian of the decoder, CMF ensures that the rules of the latent space (distances and shapes) are preserved across demographic groups. We validate CMF on synthetic Structural Causal Models (SCMs), demonstrating that it effectively disentangles sensitive geometric warping while preserving task utility, offering a rigorous quantification of the fairness-utility trade-off via geometric metrics.

---

### Flow Matching and Diffusion Models via PointNet for Generating Fluid Fields on Irregular Geometries

Authors: Ali Kashefi
Published: 2026-01-06
Categories: cs.CV, cs.LG, physics.comp-ph
arXiv: https://arxiv.org/abs/2601.03030v2

We present two novel generative geometric deep learning frameworks, termed Flow Matching PointNet and Diffusion PointNet, for predicting fluid flow variables on irregular geometries by incorporating PointNet into flow matching and diffusion models, respectively. In these frameworks, a reverse generative process reconstructs physical fields from standard Gaussian noise conditioned on unseen geometries. The proposed approaches operate directly on point-cloud representations of computational domains (e.g., grid vertices of finite-volume meshes) and therefore avoid the limitations of pixelation used to project geometries onto uniform lattices, as is common in U-Net-based flow matching and diffusion models. In contrast to graph neural network-based diffusion models, Flow Matching PointNet and Diffusion PointNet do not exhibit high-frequency noise artifacts in the predicted fields. Moreover, unlike such approaches, which require auxiliary intermediate networks to condition geometry, the proposed frameworks rely solely on PointNet, resulting in a simple and unified architecture. The performance of the proposed frameworks is evaluated on steady incompressible flow past a cylinder, using a geometric dataset constructed by varying the cylinder's cross-sectional shape and orientation across samples. The results demonstrate that Flow Matching PointNet and Diffusion PointNet achieve more accurate predictions of velocity and pressure fields, as well as lift and drag forces, and exhibit greater robustness to incomplete geometries compared to a vanilla PointNet with the same number of trainable parameters.

---

### Flow Matching and Diffusion Models via PointNet for Generating Fluid Fields on Irregular Geometries

Authors: Ali Kashefi
Published: 2026-01-06
Categories: cs.CV, cs.LG, physics.comp-ph
arXiv: https://arxiv.org/abs/2601.03030v1

We present two novel generative geometric deep learning frameworks, termed Flow Matching PointNet and Diffusion PointNet, for predicting fluid flow variables on irregular geometries by incorporating PointNet into flow matching and diffusion models, respectively. In these frameworks, a reverse generative process reconstructs physical fields from standard Gaussian noise conditioned on unseen geometries. The proposed approaches operate directly on point-cloud representations of computational domains (e.g., grid vertices of finite-volume meshes) and therefore avoid the limitations of pixelation used to project geometries onto uniform lattices. In contrast to graph neural network-based diffusion models, Flow Matching PointNet and Diffusion PointNet do not exhibit high-frequency noise artifacts in the predicted fields. Moreover, unlike such approaches, which require auxiliary intermediate networks to condition geometry, the proposed frameworks rely solely on PointNet, resulting in a simple and unified architecture. The performance of the proposed frameworks is evaluated on steady incompressible flow past a cylinder, using a geometric dataset constructed by varying the cylinder's cross-sectional shape and orientation across samples. The results demonstrate that Flow Matching PointNet and Diffusion PointNet achieve more accurate predictions of velocity and pressure fields, as well as lift and drag forces, and exhibit greater robustness to incomplete geometries compared to a vanilla PointNet with the same number of trainable parameters.

---

### Reducing Hallucinations in LLMs via Factuality-Aware Preference Learning

Authors: Sindhuja Chaduvula, Ahmed Y. Radwan, Azib Farooq, Yani Ioannou, Shaina Raza
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03027v3

Preference alignment methods such as RLHF and Direct Preference Optimization (DPO) improve instruction following, but they can also reinforce hallucinations when preference judgments reward fluency and confidence over factual correctness. We introduce F-DPO (Factuality-aware Direct Preference Optimization), a simple extension of DPO that uses only binary factuality labels. F-DPO (i) applies a label-flipping transformation that corrects misordered preference pairs so the chosen response is never less factual than the rejected one, and (ii) adds a factuality-aware margin that emphasizes pairs with clear correctness differences, while reducing to standard DPO when both responses share the same factuality. We construct factuality-aware preference data by augmenting DPO pairs with binary factuality indicators and synthetic hallucinated variants. Across seven open-weight LLMs (1B-14B), F-DPO consistently improves factuality and reduces hallucination rates relative to both base models and standard DPO. On Qwen3-8B, F-DPO reduces hallucination rates by 5x(from 0.424 to 0.084) while improving factuality scores by 50% (from 5.26 to 7.90). F-DPO also generalizes to out-of-distribution benchmarks: on TruthfulQA, Qwen2.5-14B achieves +17% MC1 accuracy (0.500 to 0.585) and +49% MC2 accuracy (0.357 to 0.531). F-DPO requires no auxiliary reward model, token-level annotations, or multi-stage training.

---

### Reducing Hallucinations in LLMs via Factuality-Aware Preference Learning

Authors: Sindhuja Chaduvula, Ahmed Y. Radwan, Azib Farooq, Yani Ioannou, Shaina Raza
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03027v1

Preference alignment methods such as RLHF and Direct Preference Optimization (DPO) improve instruction following, but they can also reinforce hallucinations when preference judgments reward fluency and confidence over factual correctness. We introduce F-DPO (Factuality-aware Direct Preference Optimization), a simple extension of DPO that uses only binary factuality labels. F-DPO (i) applies a label-flipping transformation that corrects misordered preference pairs so the chosen response is never less factual than the rejected one, and (ii) adds a factuality-aware margin that emphasizes pairs with clear correctness differences, while reducing to standard DPO when both responses share the same factuality. We construct factuality-aware preference data by augmenting DPO pairs with binary factuality indicators and synthetic hallucinated variants. Across seven open-weight LLMs (1B-14B), F-DPO consistently improves factuality and reduces hallucination rates relative to both base models and standard DPO. On Qwen3-8B, F-DPO reduces hallucination rates by five times (from 0.424 to 0.084) while improving factuality scores by 50 percent (from 5.26 to 7.90). F-DPO also generalizes to out-of-distribution benchmarks: on TruthfulQA, Qwen2.5-14B achieves plus 17 percent MC1 accuracy (0.500 to 0.585) and plus 49 percent MC2 accuracy (0.357 to 0.531). F-DPO requires no auxiliary reward model, token-level annotations, or multi-stage training.

---

### LittiChoQA: Literary Texts in Indic Languages Chosen for Question Answering

Authors: Aarya Khandelwal, Ritwik Mishra, Rajiv Ratn Shah
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03025v1

Long-context question answering (QA) over literary texts poses significant challenges for modern large language models, particularly in low-resource languages. We address the scarcity of long-context QA resources for Indic languages by introducing LittiChoQA, the largest literary QA dataset to date covering many languages spoken in the Gangetic plains of India. The dataset comprises over 270K automatically generated question-answer pairs with a balanced distribution of factoid and non-factoid questions, generated from naturally authored literary texts collected from the open web. We evaluate multiple multilingual LLMs on non-factoid, abstractive QA, under both full-context and context-shortened settings. Results demonstrate a clear trade-off between performance and efficiency: full-context fine-tuning yields the highest token-level and semantic-level scores, while context shortening substantially improves throughput. Among the evaluated models, Krutrim-2 achieves the strongest performance, obtaining a semantic score of 76.1 with full context. While, in shortened context settings it scores 74.9 with answer paragraph selection and 71.4 with vector-based retrieval. Qualitative evaluations further corroborate these findings.

---

### SA-ResGS: Self-Augmented Residual 3D Gaussian Splatting for Next Best View Selection

Authors: Kim Jun-Seong, Tae-Hyun Oh, Eduardo Pérez-Pellitero, Youngkyoon Jang
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03024v3

We propose Self-Augmented Residual 3D Gaussian Splatting (SA-ResGS), a novel framework to stabilize uncertainty quantification and enhancing uncertainty-aware supervision in next-best-view (NBV) selection for active scene reconstruction. SA-ResGS improves both the reliability of uncertainty estimates and their effectiveness for supervision by generating Self-Augmented point clouds (SA-Points) via triangulation between a training view and a rasterized extrapolated view, enabling efficient scene coverage estimation. While improving scene coverage through physically guided view selection, SA-ResGS also addresses the challenge of under-supervised Gaussians, exacerbated by sparse and wide-baseline views, by introducing the first residual learning strategy tailored for 3D Gaussian Splatting. This targeted supervision enhances gradient flow in high-uncertainty Gaussians by combining uncertainty-driven filtering with dropout- and hard-negative-mining-inspired sampling. Our contributions are threefold: (1) a physically grounded view selection strategy that promotes efficient and uniform scene coverage; (2) an uncertainty-aware residual supervision scheme that amplifies learning signals for weakly contributing Gaussians, improving training stability and uncertainty estimation across scenes with diverse camera distributions; (3) an implicit unbiasing of uncertainty quantification as a consequence of constrained view selection and residual supervision, which together mitigate conflicting effects of wide-baseline exploration and sparse-view ambiguity in NBV planning. Experiments on active view selection demonstrate that SA-ResGS outperforms state-of-the-art baselines in both reconstruction quality and view selection robustness.

---

### SA-ResGS: Self-Augmented Residual 3D Gaussian Splatting for Next Best View Selection

Authors: Kim Jun-Seong, Tae-Hyun Oh, Eduardo Pérez-Pellitero, Youngkyoon Jang
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03024v1

We propose Self-Augmented Residual 3D Gaussian Splatting (SA-ResGS), a novel framework to stabilize uncertainty quantification and enhancing uncertainty-aware supervision in next-best-view (NBV) selection for active scene reconstruction. SA-ResGS improves both the reliability of uncertainty estimates and their effectiveness for supervision by generating Self-Augmented point clouds (SA-Points) via triangulation between a training view and a rasterized extrapolated view, enabling efficient scene coverage estimation. While improving scene coverage through physically guided view selection, SA-ResGS also addresses the challenge of under-supervised Gaussians, exacerbated by sparse and wide-baseline views, by introducing the first residual learning strategy tailored for 3D Gaussian Splatting. This targeted supervision enhances gradient flow in high-uncertainty Gaussians by combining uncertainty-driven filtering with dropout- and hard-negative-mining-inspired sampling. Our contributions are threefold: (1) a physically grounded view selection strategy that promotes efficient and uniform scene coverage; (2) an uncertainty-aware residual supervision scheme that amplifies learning signals for weakly contributing Gaussians, improving training stability and uncertainty estimation across scenes with diverse camera distributions; (3) an implicit unbiasing of uncertainty quantification as a consequence of constrained view selection and residual supervision, which together mitigate conflicting effects of wide-baseline exploration and sparse-view ambiguity in NBV planning. Experiments on active view selection demonstrate that SA-ResGS outperforms state-of-the-art baselines in both reconstruction quality and view selection robustness.

---

### MedDialogRubrics: A Comprehensive Benchmark and Evaluation Framework for Multi-turn Medical Consultations in Large Language Models

Authors: Lecheng Gong, Weimin Fang, Ting Yang, Dongjie Tao, Chunxiao Guo, Peng Wei, Bo Xie, Jinqun Guan, Zixiao Chen, Fang Shi, Jinjie Gu, Junwei Liu
Published: 2026-01-06
Categories: cs.CL, cs.HC
arXiv: https://arxiv.org/abs/2601.03023v2

Medical conversational AI (AI) plays a pivotal role in the development of safer and more effective medical dialogue systems. However, existing benchmarks and evaluation frameworks for assessing the information-gathering and diagnostic reasoning abilities of medical large language models (LLMs) have not been rigorously evaluated. To address these gaps, we present MedDialogRubrics, a novel benchmark comprising 5,200 synthetically constructed patient cases and over 60,000 fine-grained evaluation rubrics generated by LLMs and subsequently refined by clinical experts, specifically designed to assess the multi-turn diagnostic capabilities of LLM. Our framework employs a multi-agent system to synthesize realistic patient records and chief complaints from underlying disease knowledge without accessing real-world electronic health records, thereby mitigating privacy and data-governance concerns. We design a robust Patient Agent that is limited to a set of atomic medical facts and augmented with a dynamic guidance mechanism that continuously detects and corrects hallucinations throughout the dialogue, ensuring internal coherence and clinical plausibility of the simulated cases. Furthermore, we propose a structured LLM-based and expert-annotated rubric-generation pipeline that retrieves Evidence-Based Medicine (EBM) guidelines and utilizes the reject sampling to derive a prioritized set of rubric items ("must-ask" items) for each case. We perform a comprehensive evaluation of state-of-the-art models and demonstrate that, across multiple assessment dimensions, current models face substantial challenges. Our results indicate that improving medical dialogue will require advances in dialogue management architectures, not just incremental tuning of the base-model.

---

### MedDialogRubrics: A Comprehensive Benchmark and Evaluation Framework for Multi-turn Medical Consultations in Large Language Models

Authors: Lecheng Gong, Weimin Fang, Ting Yang, Dongjie Tao, Chunxiao Guo, Peng Wei, Bo Xie, Jinqun Guan, Zixiao Chen, Fang Shi, Jinjie Gu, Junwei Liu
Published: 2026-01-06
Categories: cs.CL, cs.HC
arXiv: https://arxiv.org/abs/2601.03023v1

Medical conversational AI (AI) plays a pivotal role in the development of safer and more effective medical dialogue systems. However, existing benchmarks and evaluation frameworks for assessing the information-gathering and diagnostic reasoning abilities of medical large language models (LLMs) have not been rigorously evaluated. To address these gaps, we present MedDialogRubrics, a novel benchmark comprising 5,200 synthetically constructed patient cases and over 60,000 fine-grained evaluation rubrics generated by LLMs and subsequently refined by clinical experts, specifically designed to assess the multi-turn diagnostic capabilities of LLM. Our framework employs a multi-agent system to synthesize realistic patient records and chief complaints from underlying disease knowledge without accessing real-world electronic health records, thereby mitigating privacy and data-governance concerns. We design a robust Patient Agent that is limited to a set of atomic medical facts and augmented with a dynamic guidance mechanism that continuously detects and corrects hallucinations throughout the dialogue, ensuring internal coherence and clinical plausibility of the simulated cases. Furthermore, we propose a structured LLM-based and expert-annotated rubric-generation pipeline that retrieves Evidence-Based Medicine (EBM) guidelines and utilizes the reject sampling to derive a prioritized set of rubric items ("must-ask" items) for each case. We perform a comprehensive evaluation of state-of-the-art models and demonstrate that, across multiple assessment dimensions, current models face substantial challenges. Our results indicate that improving medical dialogue will require advances in dialogue management architectures, not just incremental tuning of the base-model.

---

### DNACHUNKER: Learnable Tokenization for DNA Language Models

Authors: Taewon Kim, Jihwan Shin, Hyomin Kim, Youngmok Jung, Jonghoon Lee, Won-Chul Lee, Sungsoo Ahn, Insu Han
Published: 2026-01-06
Categories: q-bio.GN, cs.CL
arXiv: https://arxiv.org/abs/2601.03019v4

DNA language models are increasingly used to represent genomic sequence, yet their effectiveness depends critically on how raw nucleotides are converted into model inputs. Unlike natural language, DNA offers no canonical boundaries, making fixed tokenizations a brittle design choice under shifts, indels, and local repeats. We introduce DNAChunker, a masked DNA language model that incorporates a learnable adaptive segmentation module to produce context-dependent, variable-length units. Building on a dynamic segmentation procedure, DNAChunker learns to allocate finer granularity to functionally enriched regions while compressing repetitive or redundant sequence. We pretrain DNAChunker on the human reference genome and evaluate it across five benchmarks, where it consistently improves over strong fixed-tokenization baselines. Further analyses and ablations indicate that unlike fixed tokenizations, segmentation is learned in a biologically-informed, mutation-resilient manner.

---