Last night, I posted a post of ChatGPT and the comparison between ChatGPT, because I referred to ChatGPT, and I found that there was no original during the day, and I knew that it was too wrong after feedback.
The screenshot is directly on the screenshot below
Corrected once. But it still tells me that Baidu Wenxin mainly uses LSTM and Transformer models for Chinese text generation tasks.
I sent it last night's mistake, and it's still telling you the wrong answer in a serious book.
Corrected the second time.
Corrected the third time, it seems to be corrected:
"Yes, Baidu has also adopted the ERNIE model for semantic understanding and text generation. ERNIE is a pre-trained language model proposed by Baidu natural language processing team, which adopts multi-task learning and knowledge distillation and other technologies, which can carry out in-depth semantic understanding and representation of text, and has a wide range of application prospects.
Baidu is developed based on the ERNIE model, using deep learning technology, training the ERNIE model through a large amount of data to achieve semantic understanding of user statements and generate natural and smooth replies. The ERNIE model is used to implement core functions such as semantic matching, answer extraction, and text generation, providing users with a more intelligent and convenient interactive experience. ”
Well, at last it goes back to square one and tells you the wrong answer.
To sum up, a lot of time was wasted, and as a result, the above mistakes were made. So if you habitually trust ChatGPT, it may be more troublesome, and the article is found to be problematic because it has been reviewed. But if you don't notice it, it will not mislead many people.