| CLIP | Contrastive Language-Image Pre-training | 对比语言图像预训练 |
| FFN | Feed-Forward Network | 前馈网络 |
| GGML | GPT-Generated Model Language | GPT生成的模型语言 |
| GGUF | GPT-Generated Unified Format | GPT生成的统一格式 |
| GQA | Grouped Query Attention | 分组查询注意力 |
| MHA | Multi-Head Attention | 多头注意力 |
| MQA | Multi-Query Attention | 多查询注意力 |
| QKV | Query, Key, Value | 查询、键、值 |
| RAG | Retrieval-Augmented Generation | 检索增强生成 |
| RTN | Reversible Transformer Network | 可逆变换网络 |
| ppl | perplexity | 困惑度 |
| MMLU | Multi-Modal Language Understanding | 多模态语言理解 |
| CEVAL | Contrastive Evaluation | 对比评估 |
| CMMLU | Cross-Modal Multi-Modal Language Understanding | 跨模态多模态语言理解 |