Technology ❯ Artificial Intelligence ❯ Natural Language Processing

Language Models

Large Language Models ChatGPT Retrieval-Augmented Generation GPT-5 Multilingual Models Multilingual Reasoning OpenAI Small Language Models Prompt Engineering Retrieval Augmented Generation Multilingual Datasets Text Generation GPT Arabic Language Models Multimodal Models Evaluation Transformer Models Benchmarking Clinical Applications Evaluation Metrics Hallucination Detection GPT-4 OpenELM DolphinGemma Prompt Injection Audio Summaries Prompt Injection Attacks Submissive Behavior Synthetic Data Generation Multi-step Reasoning Gemini Robotics SDK Gemini Nano Reasoning Mechanisms Cost-Quality Trade-offs Linguistic Diversity Decoding Processes Translation Quality Evaluation Systems Evaluation Methods Evaluation Techniques Conversational Recommendation Structured Output Bias Evaluation Logic-Augmented Generation Fundamental Language Model Applications in Software Development BharatGen Initiative Generative Models Avoidance Decoding Persuasion Techniques Historical Language Models Query Generation Optimized Prompts Readability Assessment Interactive Systems Welsh Language AI Native Language Identification Gemini 2.5 Personality Assessment Multilingual Evaluation Answer Generation Confidence Estimation Token Representations Predictive Uncertainty Conversational AI Multilingualism Gemma-based Model Temporal Reasoning Multilingual Safeguards Autoregressive Models PsyLLM Code Generation Implications for AGI Long-Context Language Models Explainability Logical Reasoning Mathematical Reasoning Bias in AI Zero-shot Learning Gemma Transformers.js AI Ethics Optical Character Recognition Ethical Implications Telecom Applications Reward Modeling Adversarial Attacks Engram Mechanism World Modeling Toxic Outputs Feedback Classification Content Generation Legal Implications Feedback Generation Nemotron-4 4B NIM HTML-to-Markdown Conversion WordLlama Model Releases AI in Elections MobileLLM

New Research Pushes Next‑Gen RAG: Process‑Level RL, Shared‑Space Retrieval, and an SAE Hallucination Detector

Most claims come from preprints and engineering posts and await broader validation.

Apple’s MLX Taps M5 Neural Accelerators for Up to 4x Faster Local LLMs

New Papers Flag Retrieval as RAG’s Weak Link, Propose Practical Fixes

Study Finds LLMs Obey Illogical Medical Requests as New Benchmarks Expose Wider Reasoning Gaps

Enterprises Scale RAG in Production as New Studies Expose Recall Gaps and Offer Targeted Fixes

Karpathy Releases Nanochat, a $100 ChatGPT‑Style LLM Speedrun, as Community Publishes Hugging Face Build

Google Begins Rolling Out Gemini for Home as Early Users Report AI-Powered Camera Summaries

New HANRAG Preprint Targets Multi‑hop Noise as RAG Guides Proliferate

UK-LLM, NVIDIA Unveil Welsh-Reasoning AI Trained on Isambard-AI

Sam Altman Says AI Talk on X and Reddit Now ‘Feels Very Fake’

University of Pennsylvania Study Shows Persuasion Can Bypass ChatGPT Guardrails on GPT-4o Mini

Developers Embrace RAG to Ground Language Models in Accurate, Up-to-Date Data

Mistral AI Unveils Europe’s First AI Reasoning Models