diff --git a/README.md b/README.md index 277bfba..ac614fd 100644 --- a/README.md +++ b/README.md @@ -24,7 +24,7 @@ ## Chinese Version -To facilitate the reading of our (English-verison) survey, we also employ LLMs + some human checking to generate a [**Chinese version**](assets/LLM_Survey__Chinese_V1.pdf) for this survey. While, since it is mainly generated by LLMs, please don't forward or post its content on the Web. +To facilitate the reading of our (English-verison) survey, we also translate a [**Chinese version**](assets/LLM_Survey_Chinese.pdf) for this survey. We will continue to update the Chinese version. @@ -663,6 +663,7 @@ Please click [here](Experiments/README.md) to view more detailed information. **Attention** 1. Multi-query attention: **"Fast Transformer Decoding: One Write-Head is All You Need"**. *Noam Shazeer*. arXiv 2019. [[paper](https://arxiv.org/abs/1911.02150)] 1. FlashAttention: **"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness"**. *Tri Dao et al*. NeurIPS 2022. [[paper](https://arxiv.org/abs/2205.14135)] +1. PagedAttention: **"vLLM: Easy, Fast, and Cheap LLM Serving with PagedAttention"**. *Woosuk Kwon et al*. 2023. paper(Stay Tuned) [[Offical WebSite](https://vllm.ai/)] ##### Analysis @@ -749,7 +750,7 @@ Please click [here](Experiments/README.md) to view more detailed information. 1. **"Scaling Laws for Reward Model Overoptimization"**. *Leo Gao et al*. arXiv 2022. [[Paper](https://arxiv.org/abs/2210.10760)] 1. **"The Wisdom of Hindsight Makes Language Models Better Instruction Followers"**. *Tianjun Zhang et al*. arXiv 2023. [[Paper](https://arxiv.org/abs/2302.05206)] 1. **"RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment"**. *Hanze Dong et al*. arXiv 2023. [[Paper](https://arxiv.org/abs/2304.06767)] - +1. **"Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment"**. *Rishabh Bhardwaj et al*. arXiv 2023. [[Paper](https://arxiv.org/abs/2308.09662)] #### Parameter-Efficient Model Adaptation 1. **"Parameter-Efficient Transfer Learning for NLP"**. *Neil Houlsby et al*. ICML 2019. [[Paper](https://arxiv.org/abs/1902.00751)] [[GitHub](https://github.com/google-research/adapter-bert)] @@ -953,6 +954,10 @@ Please click [here](Experiments/README.md) to view more detailed information. 81. **"Yes but.. Can ChatGPT Identify Entities in Historical Documents?"**. *Carlos-Emiliano González-Gallardo et al.* arXiv 2023. [[Paper](https://arxiv.org/abs/2303.17322v1)] 82. **"Uncovering ChatGPT's Capabilities in Recommender Systems"**. *Sunhao Dai et al.* arXiv 2023. [[Paper](https://arxiv.org/abs/2305.02182)] 83. **"Editing Large Language Models: Problems, Methods, and Opportunities"**. *Yunzhi Yao et al.* arXiv 2023. [[Paper](https://arxiv.org/abs/2305.13172)] +84. **"Red teaming ChatGPT via Jailbreaking: Bias, Robustness, Reliability and Toxicity"**. *Terry Yue Zhuo et al.* arXiv 2023. [[Paper](https://arxiv.org/abs/2301.12867)] +85. **"On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on Codex"**. *Terry Yue Zhuo et al.* EACL 2023. [[Paper](https://arxiv.org/abs/2301.12868)] +86. **"A Systematic Study and Comprehensive Evaluation of ChatGPT on Benchmark Datasets"**. Laskar et al.* ACL'23. [[Paper]](https://arxiv.org/abs/2305.18486) +87. **"Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment"**. *Rishabh Bhardwaj et al*. arXiv 2023. [[Paper](https://arxiv.org/abs/2308.09662)] ### The Team diff --git a/assets/LLM_Survey_Chinese.pdf b/assets/LLM_Survey_Chinese.pdf new file mode 100644 index 0000000..b8f4af3 Binary files /dev/null and b/assets/LLM_Survey_Chinese.pdf differ diff --git a/assets/LLM_Survey__Chinese_V1.pdf b/assets/LLM_Survey__Chinese_V1.pdf deleted file mode 100644 index 991a01d..0000000 Binary files a/assets/LLM_Survey__Chinese_V1.pdf and /dev/null differ diff --git a/assets/chinese_version.png b/assets/chinese_version.png index 88d7342..50bde10 100644 Binary files a/assets/chinese_version.png and b/assets/chinese_version.png differ