Bibliography (3):

1. LLaMA-2: Open Foundation and Fine-Tuned Chat Models

2. LLaMa-1: Open and Efficient Foundation Language Models

3. Averaging Weights Leads to Wider Optima and Better Generalization