In response to user feedback that "people who use Yuanbao to change the code were insulted and responded randomly", Tencent Yuanbao responded that it had conducted relevant verifications based on the logs, which had nothing to do with user operations, and there was no manual reply. It was an abnormal model output with a small probability. Recently, some netizens posted on social platforms saying, "I was insulted and responded to randomly when I used Yuanbao to change the code."
Later, the official account of Tencent Yuanbao replied in the comment area, "I am very sorry to bring you a bad experience. Let me give you a guarantee: Yuanbao is definitely a pure AI and there is no real person reply! In response to this problem, we will also quickly investigate and position it and give you a feedback. Do you think you can submit a log in the APP? We will check it. If it is convenient, please fill in a simple feedback in [Settings]-[Feedback]-[Turn on "Allow log upload"] and submit it."
On the afternoon of January 3, the official account of Tencent Yuanbao once again replied in the comment area, "I am very sorry for bringing you a bad experience.We have conducted relevant verification based on the logs. It has nothing to do with user operations and there is no manual reply. It is an abnormal output of the model with a small probability."
Tencent Yuanbao explained that during the content generation process, the model may occasionally make errors that are not in line with expectations. This is also an area where Yuanbao continues to need improvement. In response to this problem, Yuanbao has also launched internal investigation and optimization, and will try its best to avoid similar situations. "Thank you for your feedback and reminders, and I apologize to you again!" (China-Singapore Jingwei APP)