Research

In a LLM-powered autonomous agent system, LLM functions as the agent’s brain, complemented by several key components planning, memory, tool use

GPT-4, a large multimodal model.

GPT-3.5, trained language models that are much better at following user intentions than GPT-3 while also making them more truthful and less toxic.

GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model.

GPT-2, fine-tuned the 774M parameter GPT-2 language model using human feedback for various tasks.

GPT-1, a combination of two existing ideas transformers and unsupervised pre-training.

Propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely.