“Jump to ask" is an AI efficiency tool developed by Step Star based on the Step-1 and Step-1V 100 billion parameter models. It was officially opened to the public a month ago.
Step-1's performance in logical reasoning, Chinese knowledge, English knowledge, mathematics, code, etc. has surpassed GPT-3.5 in all aspects.
Step-1V ranked first in the multimodal model evaluation of OpenCompass, China's authoritative large-scale model evaluation platform (March data, currently ranked fourth), the performance is comparable to GPT-4V.
It has the ability to understand and generate text, and analyze image content, providing users with a variety of services including information query, language learning, creative writing, graphic interpretation, and one-click generation of long information images.
Yuewen can quickly and accurately analyze data, helping users quickly grasp key data; in terms of paper speed reading, it can quickly understand the core content of the paper;
The “One Picture to Understand” function is even more convenient. Users only need to enter a link or upload a document, and Yuewen will generate a long visual picture with one click, simplifying the understanding and transmission of information.
In addition, Yuewen can also read pictures and write poems, create poems using image content, and provide map signs with accompanying explanations to enhance the expression and transmission of information.
At present, in order to meet the needs of different devices and scenarios, Yuewen not only provides a web version and a public account entrance, but also specially developedMobile apps for Android and iOS (in beta)
APP
Since "Yuewen" is in the APP internal testing stage, the official will continue to iterate quickly to provide a better experience of using Yuewen on mobile phones. There may be more update prompts (including voice updates).
The main features of the Android app are as follows:
One picture to understand the function of generating long pictures
Users can quickly convert document content into a long image format that is easy to share and read by uploading files in PDF or Word format or entering a web link.
This service is particularly suitable for users who need to convert complex text materials, such as government reports, corporate financial statements or policy documents, into visual information.
Through Yuewen, they can easily transform this information into more intuitive and easy-to-understand long pictures, which are not only convenient for personal understanding, but also for sharing on social media or other platforms.
In addition, Yuewen has certain restrictions on the size of uploaded files, which must not exceed 100MB, ensuring the efficiency and practicality of the service.
Now,Users can access the website or official account, or scan the following QR code to access the one-picture-to-understand function.
Visit URL:
https://yuewen.cn/textposter
Image content analysis
Yuewen's image analysis function can process and analyze pictures uploaded by users and identify key information in the images, such as text, data, and charts.
It can create relevant texts based on image content, such as writing poems based on images or generating descriptive texts, thus enhancing the creativity and expressiveness of the content.
In addition, Yuewen can interpret the geographic location information in the image and provide users with detailed explanations and analysis of the locations in the image.
It should be noted that「Yuewen」 supports uploading images in jpg, png, jpeg and other formats, and supports up to 5 images at a time.
Information query and processing
In terms of information query, Yuewen can quickly extract key information from large amounts of data, helping users to efficiently obtain the required information.
There is no need to say much about network analysis and link parsing, which are the basic operations at present.
passUpload documents in pdf, doc/docx and other formats (up to 10 documents can be uploaded each time), through advanced natural language processing and retrieval technologies, it can accurately understand users’ questions and instructions and provide accurate and helpful answers.
In addition, it can handle complex mathematical problems, provide logical reasoning and data analysis, and help users make more informed decisions in academic research or daily work.
Yuewen is not limited to simple questions and answers, but can also provide support in creative writing, such as helping to conceive article outlines, generate creative content, or continue writing stories.
"Yuewen" public account, web version
Of course, in addition to the APP version, you can search for "跃问" in the WeChat applet or homepage search interface, jump to the official account and click on the product to try it out (there is currently no direct applet version, so you can still access it through the mobile web page)
A more convenient way is to use it directly through the URL.
Website link:
Although Yuewen provides dual-end access, the functions are the same, and no optimization or addition of functions are made on the web version (for example, Kimi's web version has more functions).
The only new feature is the ability to share the conversation by copying the relevant link and sending it to the target group.
It is understood that this function will be gradually launched on the APP, and users only need to click share to transfer it to social software such as QQ or WeChat.
One More Thing
Yuewen is a typical example of how Jieyuexingchen applies AI technology in C-end scenarios, demonstrating its leading level and innovative capabilities in large-scale model technology.
With the support of two trillion-dollar models (language and multimodal), Yuewen can accurately describe and understand the text, data, charts and other information in the image, and successfully complete multiple tasks such as content creation, logical reasoning, and data analysis.