paper reading

把看过的论文做一个总结，防止我忘掉 # 阅读习惯 1. 看到公式变量用绿色标注一下！ 2. 几遍 1. 全文通读，分清楚每句话的意义，只记录重要的话，写下问题 2. 每章提炼出重点 3. 回答以上问题

paper集合

1. FULL PARAMETER FINE-TUNING FOR LARGE LANGUAGE MODELS WITH LIMITED RESOURCES
   1. 全参数|微调
2. Towards A Unified View Of Parameter-Efficient Transfer Learning
   1. 微调|高效参数|
3. 5
4. 4
5. 3
6. LISA: Layerwise Importance Sampling for Memory-Efficient Large Language Model Fine-Tuning
7. TRANSFORMER EXPLAINER: Interactive Learning of Text-Generative Models
   1. 没啥营养，把transformer的过程搞了个可视化（也没啥用），唯一的贡献就是让我重新梳理了一遍transformer
8. Aligner: Efficient Alignment by Learning to Correct
9. A Comprehensive Survey of LLM Alignment Techniques: RLHF, RLAIF, PPO, DPO and More
   1. 综述文章，对齐
10. CLLMs: Consistency Large Language Models

看的时候遇到的问题，有的需要借助实验来解决

1. 上游 upstream是什么？
2. token长什么样子？