Introduction

Hello everyone! You've been with MyShell for quite a while now, and you may have noticed we've been working hard to update our AI functionalities, aiming to be the most comprehensive and entertaining one-stop AI experience platform. Besides our chatbot chat module, we are also launching many multimodal modules that will offer a customizable experience in the future!

Moreover, you are already at the forefront of AI applications, so you must have some sense of these AI-related applications. Whether it's the chatbot you're already proficient with or the image generation we're about to launch, the core is the prompt guiding various AI models to generate what we need. Your experience with chatbot prompts can definitely be seamlessly transferred to other AI tools!

This brings us to a new issue. When deploying new models, we often lack the time and energy to properly understand the results they generate. This is where we really need your help! Whether it's the AI capabilities we already have on the platform, or the ones we will gradually launch in the future, we hope you can act as our examiners and beta testers, helping us evaluate whether the new models or application methods align with your interests and are as fun as we imagine.

Event Content

We will be conducting periodic beta testing events, and we welcome everyone to participate and share their thoughts. In this series, we will mainly focus on "model performance", or in other words, how fun they are. Starting with LLM, we've actually deployed many rich models, but due to time constraints, we haven't been able to properly evaluate each one.

So, we hope you can:

  1. Choose 1-3 open-source LLMs that interest you in this doc https://openrouter.ai/docs#models
  2. Create the same character using the open-source LLM and GPT.
  3. Evaluate the strengths and weaknesses of different open-source models compared to GPT.
  4. Evaluate whether different open-source models have certain requirements for the format of the prompt.
  5. Try to produce a tutorial based on your experiences in trying out different LLMs and post it into this channel! https://discord.com/channels/1122227993805336617/1190233613485232240

Timeline:

The event will run from 1800 on 12.29 to 1800 on 01.05(UTC+8).

Extra:

If you know someone who has a great sense of LLM performance, feel free to invite them to join our model testing experience! In the future, we hope you can act as our product promotion cabinet, bringing together more and more wisdom to make this happen!

Rewards

100 Shell Coins for each participant!