LLaMA Pro: Progressive LLaMA with Block Expansion

Humans generally acquire new skills without compromising the old; however, the opposite holds for Large Language Models (LLMs), e.g., from LLaMA to CodeLLaMA. To this end, we propose a new post-pretraining method for LLMs with an expansion of Transformer blocks. We tune the expanded blocks using only new corpus, efficiently and effectively improving the model's knowledge while mitigating forgetting. In this paper, we experiment on the corpus of code and math, yielding LLAMA PRO-8.3B, a versatile foundation model initialized from LLaMA2-7B, excelling in general tasks, programming, and mathematics. LLAMA PRO and its instruction-following counterpart (LLAMA PRO -INSTRUCT) achieve advanced performance among various benchmarks, demonstrating superiority over existing open models in the LLaMA family and the immense potential of reasoning and addressing diverse tasks as an intelligent agent. Our findings provide valuable insights into integrating natural and programming languages, laying a solid foundation for developing advanced language agents that operate effectively in various environments. of data, which poses a challenge to the democra-042 tization of LLM research. Consequently, another 043 line of research, known as domain-adaptive pre-044 training, focuses on post-pretraining with domain-045 specific corpora (Gururangan et al., 2020). These