Large Language Models

Anthropic researchers have begun mapping the features inside Claude Sonnet, giving the first detailed look inside a production-grade large language model.

OpenAI and News Corp have agreed a multi-year global partnership giving OpenAI access to content from titles including The Wall Street Journal and The Times.

Microsoft’s guide to AI terms explains language models, grounding, RAG, orchestration and memory, transformer and diffusion models, and frontier models.

A House of Lords debate signals that changes to UK copyright law are likely as the government seeks a balance between AI innovation and creators’ rights.

OpenAI allows ChatGPT access without sign-up, sparking debate on AI accessibility. Experts weigh benefits against data privacy risks.

OpenAI has improved its Fine-Tuning API and expanded its Custom Models Program, helping developers train models that fit their own domain.

Anthropic says long context windows expose LLMs to many-shot jailbreaking, where many faux dialogues override safety training, and mitigation is hard.

Mistral AI has released Mistral Large, its most advanced model, with top-tier multilingual reasoning, a longer context window and native function calling.