Meta releases LLaMA, open weights spread

Meta's LLaMA models match much larger systems, and their weights quickly spread to the open research community.

What happened and why it matters

LLaMA-13B outperformed GPT-3 on most benchmarks. The availability of strong weights spawned an ecosystem of fine-tunes and on-device models, diffusing capability far beyond frontier labs.

Sources

• Introducing LLaMA: A foundational, 65-billion-parameter large language model ↗ Meta AI · 24 Feb 2023