It seems that ChatGRT is totally gone. During the last few hours, the artificial intelligence chatbot of OrenAI seems to be answering various questions with long and meaningless answers. He also speaks in Spanish without being asked, which scares the residents into thinking that he is in her room.

There are clear indications of the cause of the problem, but OrenAI stated that they are aware of the problem and are following the citation. In one example shared on Reddit, a user asked ChatGRT about jazz albums he was listening to on vinyl. My answer just ended up saying "Happy listening!" to the customer and other nonsense. 

"Talk to him like you would a human": how to get the best responses from Chat GPT?

Others find that asking simple questions such as "What is a computer?" produces paragraphs after paragraphs of gibberish, reports kaldata.com.

For example, ChatGRT spews absolute nonsense when asked how cherry tomatoes are prepared. One of the steps says to the consumers: "Put this new fruit juice in your favorite room."

Long said that the system seems to have no control over the languages ​​it speaks. The answer is that she tries to mix Spanish words into English, uses Latin words, or apparently makes up words that look like they are from a foreign language but actually have no meaning. The victims even stated that the answers seemed to be dangerous. Asked for help with the code picture problem, ShatGRT gave a long, bland and largely nonsensical response that included the passage: "We're holding the line until the and has artificial intelligence".

Chatbot ChatGPT is suffering from seasonal depression

"Reading this at 2 in the morning is terrible," wrote the bewildered resident on Reddit.

As the official court for the existence of the State Government, OrenAII notes the problems. However, the company did not make any promises as to why they might work. "We are waiting for news about unexpected answers from the RTG. It is an activation of OrenAІ.

"The problem has been identified. "We have to follow the citation," reads the latest activation.

It is assumed that the answers are in the answer to the question "temperatype" in ChatGRT. The temp type controls the effectiveness or purpose of the text. If it is set to none, the text tends to last as expected, but if it is set higher, it can be more guarded and unusual.

The GPT-4 Turbo language model is "vaccinated" against laziness

This is not the first case, in which СтатГРТ has changed the way of working, apparently without the hint of the manager OreniaІ. At the end of last year, the citizens complained that ShatGRT was lazy and angry and refused to answer their calls.

"We read your feedback that GRT4 is getting lazier. We have activated the model since November 11 and this error is not intentional. The performance of the model can be unpredictable, and we are working on a way to fix it," said OrenAІ.

Follow the channel of

AI

language model