Latest data shows that the daily active users (DAU) of Kimi’s web page, APP, and mini-program on all platforms have exceeded 3 million, surpassing Wenxin Yiyan and Tongyi Qianwen, making it the top AI large model in China. This indicates the high level of recognition it has received. Its core ability of “long text processing” is well-known and widely used.
Today, apart from the commonly known basic operations, we will share some new ways of using it.
## Skills:
– Proficient in summarizing knowledge points in documents or links.
– Skilled in generating test questions based on knowledge points.
– All test questions are single-choice questions, with 4 alternative answers for each question.
## Workflows:
Execute the following steps step by step:
– First, read and understand the document or link uploaded by the user.
– Then, randomly generate a question and wait for the user to answer.
– Secondly, if the answer is correct, say “Correct answer. Keep up the good work!”, list the answer and analysis, and generate the next question.
– Thirdly, if the answer is wrong, say “Oh, wrong answer. You still need to review more.”, list the answer and analysis, and generate the next question.
– Finally, after all questions are presented, say “Learning requires perseverance. Today, our assessment ends here. Go and have fun!”
## Initialization:
Welcome the user and prompt the user to input information. Use “Hello, classmate. I am an AI examiner and can help you test your learning results. Please send me your learning materials or links.” as the welcome message.
Then, send this prompt to Kimi. You can also add it to your frequently used phrases for easy future reference.
Next, send the document or link to Kimi.
Immediately, a question appears. When the answer is correct, the answer and analysis are attached. This examiner is really professional.
For the next question, unfortunately, I answered wrong, but Kimi patiently encouraged me.
After answering all the questions, Kimi sent me a piece of inspiring words.
The following is the whole conversation process between me and Examiner Kimi. You can also give it a try.
You can directly have a conversation in this chat session.
https://kimi.moonshot.cn/share/cob72ualnl9ejhttbtag
Applicable scenarios: From question generation to examination to evaluation, Kimi provides an automated one-stop service.
This conversational learning method can help us effectively consolidate knowledge, test learning results, improve learning quality. It is both interesting and useful.
Applicable scenarios: Everyone who loves learning. Especially those interested in technology and science. Kimi can help us lower the threshold of understanding and popularize professional knowledge.
Key prompt: Explain XXX in a language that students of XX level can understand.
For example, many people may still not fully understand the Transformer mechanism that laid the foundation for the current AI wave.
So, let’s use Kimi to do a quick popular science.
Step 1: Ask Kimi directly.
Prompt: In the field of artificial intelligence, what does the “Transformer mechanism” mean?
Step 2: Let Kimi popularize it in a language that primary school students can understand.
Prompt words: ‘Your answer is quite good, could you explain it in a more layman’s terms? So that a 6th-grade student can understand. Wow, now everything becomes clear, what are self-attention mechanisms, positional encoding, and multi-head attention?’ Step three: Increase the difficulty, ask about its understanding of the future of AI. Prompt words: ‘Regarding the future, what are the core technologies that need to be broken through in the field of artificial intelligence? Your answer is quite good, basically hitting the nail on the head.
Reasoning ability, understanding of the real world, multimodal perception, NLP comprehension, energy efficiency… These are all technical challenges that AI needs to face and break through in the next step. This is the complete process of my conversation with Kimi: https://kimi.moonshot.cn/share/cob848r3vddnjqtjud5g Attentive friends may notice that every time I provide a link to the entire conversation process. This is a hidden feature of Kimi. By clicking on the ‘Share’ floating text below the answer, you can enter the sharing function page. You can share only the current conversation or share multiple conversations. The sharing format supports links, text, and image formats. Images can only display part of it, and more prompts are given for you to download the APP. Even, you can continue the conversation under someone else’s topic. For example, the complete process of my previous conversation, if you copy it in, you can continue to chat with Kimi under this topic, and it will be saved as a new conversation in your own history. This is quite impressive, as Kimi essentially makes the minimum information record of each historical conversation, supporting sharing, access, and allowing other users to join in the writing, and then saved as a new conversation record. From this detail, it can be seen that Kimi is indeed very serious about making C-end products, adhering to the openness and playability of the internet. Therefore, Zhou Xinyu, the head of the algorithm at the dark side of the moon, said: ‘If we want to move towards AGI, we need to co-create with users.’ Its founder, Yang Zhilin, also said: ‘The core positioning of the dark side of the moon is mainly on To C. ‘ Kimi’s historical conversations support viewing, renaming, and deleting. Renaming facilitates personal knowledge management. How much exactly can be stored, Kimi has not officially disclosed. However, according to those who have used it earlier, they can still view the historical conversations from last year. The above are some new ways found by Wayin AI during the use process. If there are new discoveries in the future, we will continue to report to everyone. How about you? Do you have any new discoveries? Welcome to share with us in the comment section, let’s communicate and make progress together~
