P
Home

LLMs & Chatbots

LLMs & Chatbots category

4 articles

Fine-tuning GPT-2 from human preferences
News8 min

Fine-tuning GPT-2 from human preferences

In 2019, OpenAI researchers led by Daniel M. Ziegler published 'Fine-Tuning Language Models from Human Preferences,' the foundational paper that introduced reinforcement learning from human feedback (RLHF) for natural language tasks. Using GPT-2 as the base model, the team demonstrated that human preference judgments — collected through pairwise comparisons — could effectively fine-tune language models across four tasks: stylistic continuation with positive sentiment, physically descriptive continuation, and summarization on the TL;DR and CNN/Daily Mail datasets. The paper's most influential finding was that models trained with 60,000 human comparisons learned to copy whole sentences verbatim from source text, revealing how labeler heuristics can shape model behavior — a lesson that remains central to modern alignment research.

LLMs & Chatbots
Multimodal neurons in artificial neural networks
News5 min

Multimodal neurons in artificial neural networks

OpenAI discovered multimodal neurons in CLIP that respond to the same concept across visual, textual, and symbolic representations, revealing how AI models develop abstract conceptual understanding and offering new insights into both model capabilities and biases.

LLMs & Chatbots
Gemini 3.5: Google's frontier intelligence model with action capabilities
News4 min

Gemini 3.5: Google's frontier intelligence model with action capabilities

Google unveiled Gemini 3.5 at I/O 2026, a frontier AI model with native action capabilities that can navigate websites, fill forms, execute code, and manage workflows. The model powers new AI Mode in Search, deeply integrated Workspace agents, and revamped Google Finance. Google describes it as 'frontier intelligence with action,' marking a shift from conversational to agentic AI.

LLMs & Chatbots
Anthropic launches Cowork — a Claude Desktop agent that works in your files, no coding required
Review4 min

Anthropic launches Cowork — a Claude Desktop agent that works in your files, no coding required

Anthropic launched Cowork, a Claude Desktop agent for macOS that operates on user files through natural language — no coding required. Built on Claude Code's architecture but designed for knowledge workers, it can generate expense reports, organize documents, and synthesize notes. Available on Claude Pro ($20/month), it competes with OpenAI's desktop agents and Microsoft Copilot in the AI productivity space.

LLMs & Chatbots