The paper "Fair Like Us? Auditing LLM Alignment in Resource Allocation" presents a method for evaluating how large language models reason about fairness in the allocation of scarce, indivisible resources. It compares LLMs’ first‑person fairness judgments with human responses across various scenarios, finding that models tend to favor stricter fairness constraints, exhibit more self‑interested behavior, and are sensitive to framing. The study also shows that current fine‑tuning datasets struggle to align LLM judgments with human ones.
By Qishen Han, Hadi Hosseini, Joshua Kavner, Samarth Khanna, Sujoy Sikdar, Lirong Xia
EnSiTa is a trilingual multi‑domain parallel dataset and benchmark for English, Sinhala, and Tamil. It contains human post‑edited training data across seven domains and professionally translated test sets for those domains plus an additional one, all produced through a multi‑year, rigorously quality‑controlled process. The authors use EnSiTa to conduct a comprehensive study of domain‑specific machine translation across six language directions, comparing from‑scratch Transformers, pre‑trained models, and decoder‑only LLMs under various training‑data sizes, model scales, and domain settings.
By Surangika Ranathunga, Nisansa de Silva, Aloka Fernando, Kavindu Warnakulasuriya, Isuru Wijesiri, Menan Velayuthan, Charitha Rathnayaka, Thivaharan Varatharajan, Sajeevi Silva, Piumi Kandanaarachchi, Uthayasanker Thayasivam
The paper shows that chain‑of‑thought (CoT) instructions can distort evaluation of vision‑language models (VLMs) when a scorer reads answer‑label logits before the model generates a rationale. On ScienceQA, Qwen2.5‑VL‑7B’s accuracy falls from 80.76% to 45.48% under this CoT‑prefix scoring, and most predictions incorrectly pick the first option. Linear probes and free generation recover most of the lost accuracy, indicating that the answer information remains in the hidden states but is missed by the early readout. The authors explain the mismatch with vocabulary and layer diagnostics, noting that probability mass shifts toward continuation tokens while answer information stays linearly accessible in later layers. The effect varies across datasets and models, but the study demonstrates that CoT‑prefix scoring can misrepresent model knowledge unless the requested and scored outputs are aligned.
By Zeyan Li, Siyuan Qiu, Jianfeng Xu
Qwen‑Planner‑Agent is a closed‑loop AI‑for‑AI framework that enables large language models to act as both developers and participants in building advanced AI systems. The framework integrates data production, model training, and deployment through a shared action‑feedback‑verification contract, employing AI‑for‑Data, AI‑for‑Training, and AI‑driven model‑harness co‑evolution. It achieves top performance on MobilePA‑Bench by improving tool use, memory, skills, and sub‑agent coordination, while also showing gains on non‑mobile benchmarks.
By Tingyu Qu, Weigao Sun, Yuecheng Liu, Yucheng Zhao, Yi Zhu, Yifeng Ding, Qiyi Wang, Sihan Cao, Pengkun Jiao, Hanlei Xie, Xiongwei Wu, Qichao Wang, Haodong Zhang, Jiajun Liu, Yuhao Wang, Yuqing Xie, Junpeng Zhao, Long Chen, Ming Ma, Sihan Yang, Ziwang Zhao, Yanhao Jia, Liangquan Gong, Feida Zhu, Yiran Zhong, Steven Hoi
The paper introduces Self‑Play Pretraining with Zero Data, a proof‑of‑concept method that lets a model generate its own training data by searching over all computable processes using a universal Turing machine. Two models— a generator that proposes byte‑sequence programs and a learner that predicts those sequences—train together, with the generator rewarded for producing data at the learner’s frontier, creating an adaptive curriculum. Experiments show that zero‑shot performance on natural datasets scales predictably with compute, and the models exhibit in‑context learning and discover mathematical sequences during training.
By Aditya Cowsik, Kfir Dolev, Michael Y. Li, G. Bruno De Luca, Nourya Cohen, Noah D. Goodman, Yoav Levine
This study evaluates automatic speech recognition (ASR) for adolescent health communication in Twi, Dagbani, and Ewe by benchmarking five ASR systems on a Bible corpus and a domain-specific ASRH dataset, then performing supervised domain adaptation with a fine‑tuned Qwen3-ASR-0.6B model. Fine‑tuning significantly lowered word and character error rates, especially for Ewe, and the adapted model was deployed in the KasaHealth voice‑first application, which received high user approval and highlighted remaining domain gaps. The work demonstrates that in‑domain data, rather than model size or computational resources, is the primary limitation for effective ASR in these languages.
By Stephen E. Moore, Akwasi Asare, Mich-Seth Owusu, Paul Azunre, Joel Budu, Lawrence A. Adu-Gyamfi
The paper introduces Minimally Invasive Steering Vector Optimization (MISVO), a method that adjusts a frozen language model’s final hidden states by adding vectors to steer outputs toward a test‑time reward while penalizing changes using the local KL geometry of the token distribution. MISVO derives an analytic Fisher term and a suffix score‑function term, showing that the suffix term is second‑order and that Fisher surrogates match the full KL gradient to first order. Experiments on preference and code‑generation tasks with 1B–14B parameter models demonstrate that MISVO achieves the highest mean reward in most settings while maintaining diversity and coherence comparable to Best‑of‑N.
By Taha Entesari, Jingyu Zhang, Daniel Khashabi, Mahyar Fazlyab
SMILESGNN is a multimodal architecture that fuses a SMILES Transformer encoder with a GATv2 graph encoder through cross‑attention, enabling interpretable clinical toxicity predictions. The model retains an explicit graph branch, allowing GNNExplainer to identify substructures linked to toxicity. On the ClinTox dataset it achieves an AUC‑ROC of 0.987 and F1 of 0.906 with only 0.4 M parameters, while on Tox21 it attains a mean AUC‑ROC of 0.750, comparable to strong single‑modality baselines.
By Quang Minh Nguyen, Thuy Quynh Nguyen, Duc Minh Le, Ho Nhat Minh Nguyen, Thanh Long Dai Doan, Trong Nghia Nguyen
The paper introduces an acceptance protocol for sensor-coordinate and polarity binding in mechatronic commissioning, separating candidate generation from release authority. It employs a frozen four‑billion‑parameter local language model to handle requirements that a deterministic parser cannot support, and only releases plans when both facts can be derived by an external gate under a sealed grammar. Evaluation on 144 tasks showed that fabricated plans were correctly rejected, no false releases occurred among 83 releases, and incorrect user answers were identified in 169 of 431 pairings.
By Mehmet Iscan
ChunkRank is an open‑source Python library that automatically determines chunk boundaries based on a target model’s tokenizer and context window, and then selects an answer from independently produced chunk candidates. It includes a registry of 90 models from 15 providers and six answer‑selection methods, and requires only three core dependencies. Experiments show that token‑exact budgeting is important across 11 languages, and that for several datasets no content‑based ranker outperforms simply taking the first non‑empty answer due to reader abstention on chunks lacking the answer.
By Amit Nautiyal, Ayush Bhatt, Gaurav Nautiyal
The paper introduces a three-level evaluation framework—behavioral deployment, LM-head readout, and probe recoverability—to distinguish whether a language model fails a syntactic test by not encoding structure or by failing to use it. Using a trilingual control-dependency benchmark, the authors find that probe recoverability consistently exceeds LM-head readout, which in turn exceeds behavioral deployment across seven models and three languages, with the largest gap observed in Qwen3-0.6B Instruct. Layer-localized activation patching shows that instruction tuning shifts the decoded layer later, suggesting decoding favors surface shortcuts and that behavioral evaluation understates what models encode while probing alone overstates what they deploy.
By Zhenyan Lu, He Wang, Xiaohui Huang
The paper introduces Accent Analogy Guidance (AAG), a training‑free sampling technique that removes accent influence from synthetic voices in cross‑lingual zero‑shot text‑to‑speech. By subtracting an accent direction derived from the model’s own predictions, AAG improves speaker similarity while maintaining the same accent level. Experiments on four open TTS models show that AAG consistently outperforms reweighting methods, achieving higher speaker similarity scores across multiple test sets.
By Yoomee Cho, Jisun Lee
The paper introduces SVGLM, a framework that integrates scalable vector graphics (SVG) primitives into vision‑language models to enable image generation within reasoning tasks. By treating SVG both as image descriptions and text instructions, SVGLM offers a compact and interpretable method for connecting text and image reasoning. The authors provide a curated SVG‑based image editing dataset and demonstrate strong SVG generation and image‑aware reasoning performance on a mathematical benchmark.
By Sunli Chen, Ding Zhong, Ziqiao Ma, Jiaxin Liu, Zeyuan Yang, Hao Zhang, Lie Lu, Joyce Chai, Chuang Gan
MOOSEnger is a simulation‑aware AI agent framework designed for the MOOSE ecosystem, integrating an interchangeable reasoning model with domain knowledge, revised simulation artifacts, MOOSE‑specific validation, and executable solver feedback. Its generate‑check‑repair‑run workflow uses MOOSE knowledge retrieval, HIT‑aware parsing, syntax metadata, diagnostics, and revision‑controlled authoring to bind evidence to each input revision and guide bounded repair before acceptance. Across 200 prompts, MOOSEnger raises executable success from 5% to 89.5% with GPT‑5.2 and from 0% to 76.5% with Gemma 4 31B, and a ten‑case benchmark shows all generated inputs meet semantic alignment, with eight also meeting numerical‑accuracy criteria.
By Mengnan Li, Jason Miller, Zaid Abulawi, Zachary Prince, Matt Kohl, Jack M. Cavaluzzi, Guillaume Giudicelli, Casey T. Icenhour, Alexander Lindsay, Cody Permann
The paper introduces a steering‑vector‑based causal attribution framework to study how large vision‑language models (LVLMs) translate visual input into emotional narratives. By creating a specialized dataset, the authors uncover a functional decoupling in the LVLM’s three‑stage Adapt‑Aggregate‑Execute mechanism: visual emotional cues are first aggregated in middle layers via sentiment‑specific attention heads, then translated into narrative generation in deeper layers through emotion‑general pathways. Using these insights, they regulate emotional information routing to strengthen attention flow and amplify semantic activation, achieving significant performance gains on the MER‑UniBench and reducing emotional hallucinations through inference‑time intervention.
By Chengsheng Zhang, Chenghao Sun, Zhining Xie, Xinmei Tian
The paper demonstrates that language models can acquire new capabilities from post‑training data even when the training text is unrelated to the target task. Using a method called Active Taskless Distillation (ATD), the authors show that a single word from a teacher model can transfer knowledge to a student model without any target‑task examples or teacher logits. Experiments on Qwen2.5-1.5B reveal significant performance gains on HumanEval+ and improvements in scientific knowledge, commonsense reasoning, and reading comprehension across various model families.
By Ziyang Zhang, Yubin Jing, Yuanhao Zeng, Yuyao Li, Haofan Wang, Yichen Gong
The paper proposes a Diagnostic Evidence Network (DENet) that augments traditional fault‑diagnosis outputs with structured evidence, including classification, predicted characteristic frequency, and temporal localization of impulses. This evidence aligns with theoretical bearing physics and can be validated at inference time, achieving high AUROC for misclassification detection without sacrificing accuracy. A QLoRA‑adapted language model then translates DENet’s evidence into maintenance reports, markedly reducing unsupported claims.
By Yuntong Chen, Jianyu Liu, Yingqi Li, Guobin Zhao, Ziang Wang, Chao Chen, Xitian Tian, Lijiang Huang
The paper introduces the concept of Language Specific Knowledge (LSK), showing that multilingual language models can answer certain queries better when prompted in a language other than English, sometimes even in low‑resource languages. It defines a language‑selection problem and presents several baseline methods, including the authors’ LSKExtractor, to empirically demonstrate that choosing the optimal language can improve question‑answering performance across datasets covering cultural and social norms. Experiments reveal non‑intuitive mappings, such as Gemma models excelling on Chinese and Middle Eastern topics in Spanish and Qwen models performing best on authority and responsibility queries in Arabic and Chinese.
By Ishika Agarwal, Nimet Beyza Bozdag, Dilek Hakkani-T\"ur
The study evaluates large language model (LLM) graders on two computer‑science exams, testing 171 configurations of closed‑ and open‑weights models. While the best LLM configuration achieved a mean absolute error of 1.64/35—better than the 2.61/35 error between two human graders—its performance was highly sensitive to the prompt. A short "strict grader" preamble caused most open‑weight models to exceed acceptable error thresholds or stop grading entirely, whereas fine‑tuning with a single LoRA adapter restored parity with human graders and reduced sensitivity to harsh prompts.
By Ali Habibullah, Yazan Alshoibi, Mohammad Alshiekh, Salman Khan, Naeemullah Khan
PROOF is a benchmark that profiles the reliability of object-level facts in instruction-tuned language models by converting a frozen Wikidata snapshot into 18,486 English multiple-choice questions grounded in 11,779 semantic facts across 101 classes, 392 properties, and 14 domains. Each question includes an explicit "I don't know" option, a "No correct option" control, and nine controlled formulations, with 1,849 questions designed as no-correct-option traps. The study evaluates 18 open-weight model deployments on 166,374 prompts, revealing wide variability in factual accuracy, sensitivity to wording changes, and the impact of decoder perturbations.
By Andrei Chetvergov, Mikhail Solovev, Timofei Sivoraksha, Stepan Ukolov, Valeriia Kuschenko, Alexander Evseev, Sergey Bolovtsov