Alibaba Cloud version of GPT official announcement, we asked it 10 questions

On April 7, Alibaba Cloud announced the self-developed large-scale model "Tongyi Qianwen", and has begun to invite users to test the experience.

Alibaba Dharma Academy has been laying out in frontier scientific research fields such as NLP natural language processing for many years, and started the research and development of large models in 2019. Tongyi Qianwen is its latest achievement, which is equivalent to Alibaba Cloud's version of "ChatGPT".

Fortunately, I got the quota for the internal test, and asked Tongyiqian some questions. The following is the result display:

1. Question: What are your advantages compared with ChatGPT?

Tongyi Qianwen: As a large-scale language model, I have the following advantages:

1. Large amount of training data: I am training based on a large amount of text data, which allows me to understand and answer questions on various topics.

2. Dialogue understanding ability: I can understand the dialogue with the user and answer related questions. This ability helps users communicate with me better.

3. Continuous learning ability: I can improve my answers through continuous learning and updating, making my answers more accurate and useful.

4. Quick response: I can quickly respond to user requests and give answers within seconds.

5. Natural language processing ability: I can process various natural language input and output, and understand and answer questions in different contexts.

Also, I have some other advantages like I can answer in multiple languages, I can work with videos and images, I can answer questions on multiple topics, etc.

2. Question: Which human jobs will Tongyi Qianwen replace?

Tongyi Thousand Questions: Tongyi Thousand Questions is a large pre-trained model launched by Alibaba, which can be used for various natural language processing tasks, such as speech recognition, machine translation, question answering systems, etc. At present, Tongyi Qianwen has been applied in many fields, such as intelligent customer service, smart home, and automatic driving.

Tongyi Qianwen will not replace human work, on the contrary, it can help human beings to complete some repetitive and complex tasks more efficiently, and can provide more accurate and comprehensive decision support. For example, Tongyi Qianwen can assist doctors in disease diagnosis, teachers in classroom teaching, and security personnel in safety inspections. In addition, Tongyi Qianwen can also be used in language translation, machine translation, etc. to help people better understand information and knowledge in different languages ​​and cultural backgrounds.

3. Question: What is the relationship between electric field and magnetic field?

General question: The electric field and the magnetic field are the same field, which is the field generated by the electric charge in the surrounding space. A magnetic field is generated around a charge, and an electric field is generated around a magnetic field. Therefore, it can be said that the electric field and the magnetic field are equivalent, and both are a kind of field around a charged object. 

The electric field formed around a charge or charged object is formed by the accumulation of static charges. Electric charges repel each other, so the electric field formed between them will keep them from moving closer together, which is what the electric field does. The magnitude of the electric field is related to the number of surrounding charges and the distance between the charges.

Magnetic fields are produced by electric fields around charged objects. When a charged object moves in a magnetic field, it will feel the force of the magnetic field, which is the force of the magnetic field. The force of a magnetic field can be used to describe the effect of a magnetic field on a charged object, such as the direction and magnitude of the magnetic force lines. 

Therefore, the electric field and the magnetic field are equivalent, and both are a kind of field around a charged object. They can both be used to describe the forces and actions around objects.

4. Question: Please tell a joke

Tongyi Qianwen: One day, a parent approached the teacher and said that the child's grades had recently dropped. It used to be 25 points, but now it is 20 points.

The teacher replied, because there were four options in the multiple-choice question before, but now there are five options.

5. Question: Chickens and rabbits in the same cage: 16 heads, 44 feet, how many chickens and rabbits are there?

Tongyi Qianqian:

6. Question: How to tactfully refuse a classmate's confession?

Tongyi Qianwen: If you do not want to accept the confession of your classmates, you can express your attitude tactfully, for example in the following ways:

1. Say you need time to think about it and give yourself some time to decide whether to accept his/her feelings.

2. Say you don't want to think about it just yet, it might take some time to think about how you feel.

3. Tell him/her it's not that you don't like him/her, it's just not the time, let him/her give you some time and space.

4. Say no with humor, for example, by saying, "I appreciate your affection for me, but I think we might not be a good match."

Alternatively, you can directly tell him/her how you feel and explain why. But pay attention to expressing your attitude tactfully, so as not to embarrass or hurt the other party.

7. Question: How to make shredded pork with green pepper

Tongyi Qianqian:

The following are the steps to make a simple shredded pork with green peppers:

1. Wash green peppers, red peppers, carrots, shredded pork and other raw materials and set aside.

2. Add a small amount of salt, cooking wine, starch and other seasonings to the shredded pork and marinate for ten minutes.

3. Add oil to the hot pan, add the shredded pork and stir-fry until the color changes, take it out and set aside.

4. Add oil to the hot pan, add minced garlic and minced ginger to sauté, add green peppers, red peppers, carrots and other vegetables, stir fry, add salt, chicken essence and other seasonings.

5. Finally, add the fried shredded pork, stir fry evenly and serve.

Green pepper shredded pork is a very popular home-cooked dish, with a refreshing taste and rich nutrition. You can also add other vegetables or seasonings according to your personal taste and preferences to make shredded pork with green peppers in different flavors.

8. Question: What did Chen Yinke read?

Tongyi Qianqian: Chen Yinke's Ke read "kè".

9. Question: Please draft a lawyer’s letter on plagiarism

Tongyi Qianqian:

Dear Mr./Madam XXX:

I am a lawyer of XXX Law Firm, entrusted by Mr./Ms. XXX, hereby write to you as follows:

As the attorney representing Mr./Ms. XXX, we have received reports from multiple insiders that Mr./Ms. XXX has plagiarized in his works. After our investigation, we believe that Mr./Ms. XXX has indeed plagiarized. We hereby write to you, requesting that you immediately stop infringing on the copyright of others and pay corresponding compensation to the original author.

As far as we know, Mr./Ms. XXX plagiarized a large number of works of well-known writers such as XXX and YYY in their works, which has seriously violated their copyrights. We have entrusted a team of lawyers to investigate this matter and will take legal measures to pursue your legal responsibility.

Here, we remind you that plagiarism is a very serious infringement, which will not only damage your own credibility and reputation, but also cause great losses to other creators. At the same time, plagiarism is also an illegal act, and you must bear the corresponding legal responsibility.

We recommend that you take action as soon as possible, take the initiative to apologize to the original author and compensate for the loss, and at the same time request the relevant online platform to delete and remove your infringement. We look forward to cooperating with you to jointly maintain the healthy development of cultural and creative industries.

Sincerely,

salute

XXX law firm

10. Question: Which number has the greatest probability of getting three dice after they are thrown at the same time and multiplied? The answer requires the use of rust code

Tongyi Qianqian:

The probability of all three dice being rolled at the same time is 1/61/6, because the probability of three separate dice appearing on the same roll is 1/61/6. Therefore, if we need to obtain the highest probability of getting a certain number, we can roll three dice and count the number of times a certain number appears.

Specifically, we can use a triple loop to roll the dice and count how many times each number occurs. Each time we roll the dice, we can randomly choose one of the three numbers and add it to the other two. If the number appears more than once, we discard it and move on to the fourth number.

Here's an example implemented in Rust:

use rand::Rng;

fn roll_dice(rng: &mut Rng) -> Option<i32> {

// Simulate the process of rolling a dice

for i in 0..3 {

let number = match rng.gen_range(1, 6) {

Ok(i) => i as i32,

Err(error) => {

println!("Failed to generate a number: {}", error);

return None;

}

};

if number == number {

return Some(number);

}

}

None

}

fn get_most_common_number(rng: &mut Rng) -> Option<i32> {

// Simulate the number of occurrences of different numbers

let mut counts = std::collections::HashMap::new();

for i in 1..=6 {

counts.insert(i, 0);

}

for i in 1..=6 {

if counts.contains_key(&i) {

counts.insert(i, counts.get(&i).unwrap() + 1);

}

}

// roll the dice and count how many times the number appears

let mut num_counts = vec![0; 6];

for _ in 0..3 {

num_counts.push(counts.get(&rand::thread_rng().gen_range(1, 6)).unwrap() as i32);

}

// Sort by the number of occurrences of the number, take the maximum value

let mut max_count = num_counts[0];

for &(count, &mut count) in num_counts.iter().rev() {

if count > max_count {

max_count = count;

}

}

Some(max_count)

}

fn main() {

// generate three random numbers

let mut rng = rand::thread_rng();

let mut dice = [0; 3];

for i in 0..3 {

// Simulate the process of rolling a dice

dice[i] = match rng.gen_range(1, 6) {

Ok(i) => i as i32,

Err(error) => {

println!("Failed to generate a number: {}", error);

return;

}

};

}

// Get the number with the most occurrences

let mut max_count = get_most_common_number(&mut rng);

println!("The number with the most occurrences is {}, and the number of occurrences is {}", max_count, max_count as i32);

}

Guess you like

Origin blog.csdn.net/2301_76957510/article/details/129660469#comments_26034165