The paper revisits the debate between Pinker & Prince (1988) and Rumelhart & McClelland (1986) regarding neural network models of English past tense. It argues that modern Encoder-Decoder architectures in NLP address the empirical shortcomings highlighted by Pinker and Prince, eliminating the need to simplify the past tense mapping problem. The authors suggest that the strong performance of these contemporary models merits a reassessment of their role in linguistic and cognitive modeling.
By Christo Kirov, Ryan Cotterell
SelfGraphRAG is a framework that generates synthetic question‑answer pairs directly from the structure of a knowledge graph to train a query‑conditioned graph retriever. By capturing multi‑hop paths and local neighborhoods, the generated questions provide relational supervision without requiring manually labeled data. Experiments on multi‑hop question answering and classification tasks show that SelfGraphRAG improves retrieval precision and downstream reasoning performance compared to embedding‑based baselines.
By Ben Lagnese, Manas Gaur
The paper studies how transformer representations evolve across layers by examining the intrinsic dimensionality (ID) of token embeddings and their neighborhood structures. It finds that closed‑class tokens expand and collapse earlier than open‑class tokens, and that these changes are linked to shifts in local geometry. The authors compare encoder and decoder models, showing distinct layer‑wise behaviors, and demonstrate that geometric features alone can predict a token’s part‑of‑speech and reveal how semantic content changes across layers.
By Samuele Vallisa, Federico Ravenda, Claudio Palominos, Rui He, Andrea Raballo, Antonietta Mira, Philipp Homan, Wolfram Hinzen
ReliableRAG is a new framework for Retrieval-Augmented Generation that tackles misinformation in multi‑hop question answering. It extracts structured triples from retrieved documents, evaluates each triple’s reliability by combining semantic relevance to the query with credibility, and keeps only the top‑K reliable, non‑redundant triples. Using these refined triples, the system builds robust reasoning chains that filter out deceptive misinformation and produce accurate, trustworthy answers.
By Jinpu Jiang, Xuan Wu, Wenhao Song, Bo Yang, You Zhou, Hongwei Ge, Heow Pueh Lee, Yanchun Liang, Chunguo Wu
The paper introduces KESA, a knowledge‑enhanced approach for sentence‑level sentiment analysis that incorporates sentiment knowledge through two auxiliary tasks: sentiment word cloze and conditional sentiment prediction. These tasks use prior sentiment polarity to guide the selection of sentiment words and the prediction of overall sentiment, respectively, and explore label combination methods to unify multiple label types. Experiments show that KESA consistently outperforms pre‑trained models and complements existing knowledge‑enhanced post‑training methods.
By Qinghua Zhao, Shuai Ma, Shuo Ren
BRIDLE is a self‑supervised encoder pretraining framework that extends bidirectional training to audio, image, and video by incorporating residual quantization (RQ) with multiple hierarchical codebooks. This approach allows fine‑grained discretization of latent representations and interleaves training between the encoder and tokenizer. Experiments show that BRIDLE achieves state‑of‑the‑art results on audio classification benchmarks and competitive performance on image and video classification tasks, outperforming traditional vector‑quantization methods.
By Hoang M. Nguyen, Satya N. Shukla, Qiang Zhang, Hanchao Yu, Sreya D. Roy, Dipesh Tamboli, Taipeng Tian, Lingjiong Zhu, Yuchen Liu
The paper introduces CuCu, a multi‑agent LLM framework that converts national social studies curricula into open‑ended, culture‑specific question‑answer pairs for fine‑tuning language models. Using the Korean curriculum, the authors build KCaQA, a dataset of 34.1k QA pairs that cover culture‑specific topics and ground responses in local sociocultural contexts. Experiments show that fine‑tuning with KCaQA improves the model’s cultural alignment and relevance to Korean society.
By Haneul Yoo, Won Ik Cho, Geunhye Kim, Jiyoon Han
The paper investigates whether retrieval‑augmented generation (RAG) can uniformly correct factual errors in large language models (LLMs) by testing six LLMs on a benchmark of about 2,000 public companies. In a controlled factual QA setting, the authors evaluate the models under four conditions—no‑context, perfect context, misleading context, and distraction context—across four atomic attributes. Results show significant geographic disparities in baseline accuracy, and while perfect context improves performance, it does not eliminate these gaps; misleading context often leads models to copy incorrect information, and larger models only marginally reduce structural biases.
By Abhinav Havaldar, Enrico Santus
The paper introduces Gavel, a framework for evaluating large language models (LLMs) on long-context legal summarization tasks. Gavel includes a reference-based component (Gavel-Ref) with checklist, residual-fact, and writing-style checks, and a reference-free component (Gavel-Agent) that assesses factual coverage directly from source documents. Experiments on 12 frontier LLMs reveal that models tend to omit key information more than hallucinate, perform well on simple checklist items but struggle with rare, complex items, and their performance degrades with longer cases. Gavel-Agent cuts token usage by at least 36% compared to traditional methods while maintaining competitive accuracy, and it also generalizes effectively to the medical domain.
By Yao Dou, Benjamin Mamut, Wei Xu
BanglaMamba explores Mamba-based State Space Models (SSMs) as a computationally efficient alternative for Bangla fake news detection. Compared to BanglaBERT and a custom BERT trained from scratch, BanglaMamba achieves a Macro‑F1 score of 0.9029, close to the 0.9057 of the custom BERT, while delivering 2.2× higher inference throughput and 49% lower peak GPU memory usage. Cross‑dataset evaluation shows BanglaBERT generalizes better, underscoring the value of large‑scale pretraining.
By M. K. Khalidi Siam
In 2026, we held the fourth iteration of the SHROOM Shared Task series: SHROOM-Visions (\textbf{S}hared-task on \textbf{H}allucinations and \textbf{R}elated \textbf{O}bservable \textbf{O}vergeneration...
Volumetric medical VQA requires reasoning over long and redundant 3D visual token sequences, especially in multi-sequence MRI where complementary modalities provide diverse diagnostic cues but expose...
Large language models (LLMs) demonstrate impressive performance across a wide range of general NLP tasks; however, their effectiveness in sensitive domains, such as hate speech detection, remains less...
The paper introduces Tlow, a flow-based item tokenizer that transforms raw semantic embeddings into a latent space following a standard normal distribution, enabling independent tokenization and simplifying distributional complexity. Tlow incorporates codebook guidance to align token embeddings with the codebook space, producing semantically clear token IDs. Experiments on four public datasets and an online multi‑modal retrieval task on WeChat show that Tlow improves recommendation performance and increases user click‑through rates by over 10%.
By Nian Li, Chonggang Song, Jingtao Ding, Lingling Yi, Yong Li, Qingmin Liao
PARTAB is a framework that improves large language model reasoning on tables by constructing a structured evidence interface. It represents query‑relevant evidence as semantically coherent, row‑linked table regions and performs hierarchical selection over column groups and row‑level partitions before composing the evidence for answer generation. Evaluations on multiple table reasoning benchmarks show that PARTAB consistently outperforms full‑table prompting and recent methods, achieving strong performance on WikiTableQuestions and TabFact while remaining competitive on numerical reasoning tasks.
By Md Mahadi Hasan Nahid, Davood Rafiei
The paper introduces Constrained Entity Selection under Partial Knowledge (CES-PK), a framework for improving large language model (LLM) based knowledge graph question answering (KGQA) by filtering candidate answers with lightweight symbolic constraints instead of full semantic parsing. CES-PK uses a three-valued constraint semantics—satisfied, violated, unknown—to handle incomplete knowledge graphs and avoid incorrect rejections under open‑world assumptions. Experiments on the Hetionet biomedical knowledge graph show that applying type, relation, and exclusion constraints increases precision while preserving recall, and that satisfied constraints can be used to rank remaining candidates.
By Emanuel Kitzelmann
SonarLLM is a multimodal large language model that treats sonar as a native perceptual modality, combining a sonar‑specific encoder, physics‑aware feature enhancement, and reliability‑aware hierarchical fusion to align acoustic structure with optical semantics. The authors introduce SonarBench, a benchmark covering recognition, counting, visual question answering, and captioning across sonar‑only, optical‑only, and fusion settings, enabling controlled measurement of cross‑modal complementarity. SonarLLM achieves 72.0% macro accuracy on sonar‑only tasks and 68.7% under fusion, outperforming baselines by significant margins and demonstrating increasing fusion gains as optical visibility degrades.
By Cong Su, longxuan ma, Ling Dong, Guofeng Tang, Weijie Yin, Haohui Chen, Zhengtao Yu
The paper introduces Gated Activation Steering, an inference-time intervention that jointly mitigates hallucination and sycophancy in medical question answering. By learning separate steering directions from contrastive clinical pairs and applying them to specific attention heads, the method uses behavior‑specific gates to intervene only when needed. Experiments on EHR‑based clinical questions show that the 4‑billion‑parameter model with gated steering outperforms its unsteered counterpart and rivals larger models in resisting user pressure.
By Himanshu Tripathi, Subash Neupane, Shaswata Mitra, Sudip Mittal, Noorbakhsh Amiri Golilarz, Shahram Rahimi
The Transformer Accelerator (TFA) is a synthesizable, parameterizable INT8 memory‑to‑memory engine designed for transformer inference and machine translation. It features a one‑time‑multiplexed datapath that handles prompt processing and autoregressive generation, and implements key operations such as matrix multiplication, softmax, RMSNorm, and elementwise functions through eight 512‑bit macro‑op descriptors. In extensive verification, TFA achieved zero mismatches across 25 tests and 34 constrained‑random runs, matched floating‑point references on multiple translation tasks, and delivered a 20× speedup over a 22‑thread CPU while projecting significant energy reductions in larger designs.
By Shashank
arXiv:2608.23864v1 Announce Type: new
Abstract: Visual tokenizers increasingly inject semantic supervision into latent spaces to make downstream diffusion easier. Yet how these semantics should be or...
By Junqiu Yu, Pandeng Li, Yikai Wang, Jiaxing Zhao, Yujie Wei, Kaixun Jiang, Quanhao Li, Hongtao Yu, Zhihang Liu, Zhaohe Liao, Junjie Zhou, Yun Zheng, Yu Liu, Yanwei Fu