Update README.md

Author: ZrrSkywalker

README.md

@@ -5,18 +5,16 @@ Official implementation of ['LLaMA-Adapter: Efficient Fine-tuning of Language Mo
 <p align="center">                                                                                                                                          <img src="docs/logo_v3.png"/ width="70%"> <br>
 </p>
 
-This repo proposes **LLaMA-Adapter**, a lightweight adaption method for fine-tuning instruction-following [LLaMA](https://github.com/facebookresearch/llama) models 🔥, using 52K data provided by [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca).
+This repo proposes **LLaMA-Adapter (V2)**, a lightweight adaption method for fine-tuning **Instruction-following** and **Multi-modal** [LLaMA](https://github.com/facebookresearch/llama) models 🔥.
 
 Try out the web demo 🤗 of LLaMA-Adapter: [![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces/csuhan/LLaMA-Adapter)
 
 ## News
-- **[2023.04.30]** We are noticed that GPT-4 score has a strong positional bias in favor of the first response. The paper will be updated to reveal this position bias. Great thanks to [Canwen Xu](https://scholar.google.com/citations?user=oopKCDMAAAAJ&hl=en).
-- **[2023.04.30]** The technical report for **LLaMA-Adapter V2** is released at [Preprint](https://arxiv.org/abs/2304.15010).
-- **[2023.04.28]** 🔥🔥 The code and model checkpoint for **LLaMA-Adapter-V2 (65B)** are NOW available [here](https://github.com/ZrrSkywalker/LLaMA-Adapter/tree/main/llama_adapter_v2_chat65b)! Cheers!!
-- **[2023.04.22]** We have released **LLaMA-Adapter V2**, a multi-modal instruction model! Check out our [demos](#demos)!
+- **[2023.04.30]** We are noticed that GPT-4 evaluation has a strong positional bias in favor of the first response. We will soon update the paper to reveal the position bias. Our approach also shows a potential to automate chatbot assessment, but still requires further research. Great thanks to [Canwen Xu](https://scholar.google.com/citations?user=oopKCDMAAAAJ&hl=en).
+- **[2023.04.30]** The technical report for **LLaMA-Adapter V2** is released at [preprint](https://arxiv.org/abs/2304.15010).
+- **[2023.04.28]** 🔥🔥 We release **LLaMA-Adapter V2 (65B)**, a multi-modal instruction model! Check out our [demos](#demos) and [code](https://github.com/ZrrSkywalker/LLaMA-Adapter/tree/main/llama_adapter_v2_chat65b)!
 - **[2023.04.15]** The **Training Code** for LLaMA-Adapter (7B) can now be found in [alpaca finetune v1](https://github.com/ZrrSkywalker/LLaMA-Adapter/tree/main/alpaca_finetuning_v1). 📌
-- **[2023.03.28]**  The technical report for LLaMA-Adapter is released at [Preprint](https://arxiv.org/pdf/2303.16199.pdf).
-- **[2023.03.28]** The generation code for LLaMA-Adapter based on 7B LLaMA has been released.
+- **[2023.03.28]**  The [technical report](https://arxiv.org/pdf/2303.16199.pdf) and generation code for LLaMA-Adapter is released.
 
 ## <div id="demos">Demos (LLaMA-Adapter V2)</div>