← Zurück zur Forschung
Language Models are Few-Shot Learners
Brown et al. • 2020
LLMIn-Context LearningGPT-3
Abstract
This paper introduced GPT-3 and the concept of in-context learning. It demonstrated that scaling language models allows them to perform tasks given only a few examples in the prompt, without gradient updates or fine-tuning.
Warum Es Wichtig Ist
- Defined the prompt engineering paradigm
- Validated scaling laws empirically
- Shifted focus toward general-purpose models
Fragen zu diesem Artikel stellen
Loading chat...
Language Models are Few-Shot Learners
Brown et al. • 2020
LLMIn-Context LearningGPT-3
Abstract
This paper introduced GPT-3 and the concept of in-context learning. It demonstrated that scaling language models allows them to perform tasks given only a few examples in the prompt, without gradient updates or fine-tuning.
Warum Es Wichtig Ist
- Defined the prompt engineering paradigm
- Validated scaling laws empirically
- Shifted focus toward general-purpose models
Fragen zu diesem Artikel stellen
Loading chat...