RUCAIBox/LLMSurvey

Curates a structured taxonomy of LLM research papers organized by development stages—pre-training, adaptation tuning, and utilization—alongside evolutionary graphs tracking model lineages (GPT-series, LLaMA family) and empirical benchmarks for instruction tuning effectiveness. Provides complementary resources including prompt engineering guidelines, ability evaluation experiments across representative tasks, and a timeline of major LLM releases to contextualize research contributions within the field's rapid evolution.

12,119 stars. No commits in the last 6 months.