GLM-4 is a Chinese-English bilingual AI model from Tsinghua University / Zhipu AI, optimized from the ground up for both languages, unlike most English-centric models. It features a Mixture of Experts architecture for fast inference, long context up to 128K tokens, and a focus on function calling and agent workflows.
This content explores the concept of semantic search, emphasizing that it begins with the use of embeddings. It delves into the technical foundation behind meaning-based information retrieval.
This article delves into the limitations of LSTMs in maintaining context, even with their improved memory capabilities over vanilla RNNs. The author uses a personal experience of learning English to illustrate the three specific problems LSTMs still don't solve, setting the stage for discussing attention mechanisms.
One year post-launch, AI Mode is changing U.S. search behavior as users increasingly shift from traditional keyword-based queries to natural language input. This indicates a significant transformation in how people interact with search engines.
This content explores the fascinating area of translating the internal processes or "thoughts" of an artificial intelligence model like Claude into understandable language. It investigates how the complex operations of AI can be interpreted and expressed to better understand its reasoning.
The author built an AI customer support system, Nova, which achieved significant success by focusing on learning from its mistakes rather than solely on minimizing errors. This approach, termed "Error Genome," led to a 40% reduction in error rates and a 20% increase in overall system accuracy.
Amazon QuickSight now generates complete multi-sheet dashboards from natural language prompts. This feature allows BI professionals to create production-ready analyses in minutes, significantly reducing manual setup time.
This research introduces a novel approach to neural language correction leveraging character-based attention mechanisms. The method aims to improve the accuracy and robustness of automatically correcting grammatical and spelling errors in text.
This paper introduces the Weighted Syntactic and Semantic Context Assessment Summary (wSSAS), a deterministic framework to optimize text categorization using LLMs. It addresses LLM limitations by organizing text hierarchically and employing a Signal-to-Noise Ratio (SNR) to focus on high-value semantic features.
This paper introduces XHS-SCoRE, a reader-grounded benchmark for detecting if a text-only Xiaohongshu (RedNote) post elicits upward, downward, or neutral social comparison. The study finds a consistent mismatch between LLM generation fluency and reliable detection ability, indicating that LLMs generate social-comparison triggers they fail to robustly detect.
This paper investigates the local geometry of embedding clouds induced by controlled classes of semantically close sentences. The authors introduce a local geometric modeling scheme and a latent probing procedure for representation-space analysis and local manifold modeling.
Este estudo apresenta o dataset TR-EduVSum, focado em vídeos educacionais turcos, e propõe o método AutoMUP. Este método gera resumos padrão-ouro de forma automática e reproduzível a partir de múltiplos resumos humanos, usando agrupamento de unidades de significado e modelagem estatística de consenso.
This work argues that observed effects from "counterfactual prompting" in LLMs cannot be attributed to a targeted factor without accounting for meaning-preserving text modifications that establish general model sensitivity. The research shows that prediction flip rates when surgically changing patient gender are statistically indistinguishable from rates induced by simply paraphrasing inputs, suggesting that special sensitivity to patient gender cannot be concluded.
This paper introduces a highly efficient Retrieval-Augmented Generation (RAG) system specifically for Ukrainian document question answering, which achieved 2nd place in the UNLP 2026 Shared Task. It features a custom hybrid search and a specialized Ukrainian language model, compressed for high-quality, verifiable local deployment on resource-constrained hardware.
Este artigo introduz o Text2DistBench, um novo benchmark para avaliar a capacidade de LLMs inferirem conhecimento distribucional a partir de linguagem natural. Diferente dos benchmarks tradicionais, ele foca em tarefas do mundo real, como estimar proporções de sentimentos ou identificar tópicos frequentes em coleções de texto como comentários do YouTube.
This paper introduces MATH-PT, a novel dataset of 1,729 mathematical problems in European and Brazilian Portuguese, to address the linguistic bias in LLM mathematical reasoning evaluations. The benchmark reveals that frontier reasoning models achieve strong performance in multiple-choice questions but their performance decreases for open-ended questions.
This paper introduces BatteryPass-12K, the first public dataset for the novel task of digital battery passport (DBP) conformance classification, addressing a critical need before new EU regulations. It benchmarks 22 language models, finding that "Thinking models" like GPT-5.4 achieve the best performance, and few-shot examples significantly enhance results on this challenging task.
This paper introduces a proactive EMR assistant for doctor-patient dialogue, designed to overcome limitations of passive systems by integrating streaming ASR, belief stabilization, and action planning. The system was evaluated in a preliminary controlled setting, achieving an F1 of 0.84 and Recall@5 of 0.87.
CogRAG+ is a training-free framework designed to diagnose and remediate memory and reasoning deficiencies in large language models for professional exam QA. It decouples and aligns retrieval and reasoning with human cognitive hierarchies, employing Reinforced Retrieval and cognition-stratified Constrained Reasoning to enhance accuracy and consistency.
This research paper addresses optimal question selection for information gathering in psychiatric intake using conversational AI. It introduces a benchmark with 655 questions and synthetic vignettes, evaluating LLM-guided adaptive policies.