Research Papers

Language Models are Few-Shot Learners

15000
Tom Brown (OpenAI), Benjamin Mann (OpenAI)

We demonstrate that scaling up language models greatly improves task-agnostic, few-shot performance...

2020-05-28 NeurIPS 2020 Language Models Few-Shot Learning GPT-3
PDF arXiv

Constitutional AI: A Framework for Machine Learning Systems that Respect Human Values

500
Anthropic Research Team (Anthropic)

We present a framework for training AI systems to respect human values and ethical principles...

2023-12-01 arXiv AI Safety Ethics Constitutional AI
PDF arXiv