大语言模型 (LLM)
定义
大语言模型是基于 Transformer 的模型，在大规模文本（有时是多模态）数据上训练。 They exhibit emergent abilities: few-shot learning, 推理, and tool use when scaled and aligned (例如 via RLHF).
一个有用的心智模型：预训练在巨大的语料库上学习下一个 token 预测，赋予模型广泛的知识and language ability. Instruction tuning (and similar) trains the model to follow user instructions and formats. Alignment (例如 RLHF, DPO) shapes behavior to be helpful, honest, and safe. At inference time you can use the model zero-shot, few-shot, or augment it with 检索 (RAG) or tools (agents).
工作原理
Pretraining learns 下一个 token 预测 on large corpora and produces a base model. Optional fine-tuning (例如 fine-tuning) adapts it to tasks or instruction formats; alignment (例如 RLHF, DPO) optimizes human preference and safety. The deployed model is then used at inference time. You can call it zero-shot (no examples), few-shot (with prompt engineering), or augment it with RAG (检索 as context) or agents (tools and loops). The diagram summarizes the training pipeline and the two main inference augmentations.
应用场景
LLMs are used wherever you need flexible language understanding or generation, from chat to code to analysis.
- Chat, summarization, and translation
- Code assistance and generation
- Question answering and research assistance (often with RAG or tools)
优缺点
| Pros | Cons |
|---|---|
| Flexible, one model for many tasks | Cost and latency |
| Strong few-shot performance | Hallucination, bias |
| Enables agents and tool use | Requires careful evaluation |
外部文档
- OpenAI – Models overview — GPT and capabilities
- Google AI for Developers — Gemini and APIs
- Anthropic – Models — Claude and documentation
- Hugging Face – NLP course — From transformers to LLMs