【中英字幕】LLM 应用开发全栈指南 | FSDL LLM Bootcamp -

本讲相关学习资料：

Robert Huben, “How does GPT-3 spend its 175B parameters?” - https://aizi.substack.com/p/how-does-gpt-3-spend-its-175b-parameters

Anthropic, “In-context Learning and Induction Heads” 深入探索了大语言模型 in-context learning 能力的来源 - https://transformer-circuits.pub/2022/in-context-learning-and-induction-heads/index.html

最近的 RedPajama 项目中尝试“复现”了LLaMA的训练数据集 - https://together.ai/blog/redpajama

Yao Fu, “How does GPT Obtain its Ability? Tracing Emergent Abilities of Language Models to their Sources” 为何要在训练中包括代码数据， GPT 模型家族谱系图， alignment tax 等内容 - https://yaofu.notion.site/How-does-GPT-Obtain-its-Ability-Tracing-Emergent-Abilities-of-Language-Models-to-their-Sources-b9a57ac0fcf74f30a1ab9e3e36fa1dc1

Open Assistant数据集 https://huggingface.co/datasets/OpenAssistant/oasst1

Anthropic: Constitutional AI https://www.anthropic.com/index/claudes-constitution

OPT优化的血泪史 https://arxiv.org/pdf/2205.01068.pdf

模型inference优化的手段 https://lilianweng.github.io/posts/2023-01-10-inference-optimization/