Bibliography (3):

1. GPT-3: Language Models are Few-Shot Learners

2. https://openai.com/index/gpt-4-research/

3. RoBERTa: A Robustly Optimized BERT Pretraining Approach