Topic-3-ChatGPT & Prompt Engineering



SEMANTICS

What is semantics in artificial intelligence?

What is Semantic AI? Semantic AI, which is also related to natural language processing (NLP) or natural language understanding, is a branch of artificial intelligence focusing on how computers understand and process human language.0

 

Idea is that you have large amount of data. Old method was that when a question was asked you start to match their words.  But as we discussed in Stemming and Lemmatization; there cane be difference in words.

Applying the tokenization technique, that is breaking the sentences in small words. Magic happened here that we converted all our data in specific numeric form called vectors and store in vector data base. Whenever a question arises, it is also converted in vectors and store in data base. So Semantic search means, searching such words which are same in meaning in context of meaning. So, when I take a query and match it with my data base having same meaning as in query, it is called Semantic Search.

Same has been done in Irfan ChatGpt, the data given by the team and data got through brain storming and after spending time with Sir Irfan was store in a modal. When a query will be asked from Irfan Chat Gpt it would be semantically matched and matching outcome to the query would come as an output.

Question arises where is Chat Gpt in all this scenario. After matching the query with similar data in database, it is ChatGpt that arranges these matching words into meaningful sentences.

Now we continue the demo of Irfan ChatGpt.










Chat wit ChatGpt can be elaborated with the help of an example. Suppose you are sitting with an old man who have lot of knowledge and wisdom. Due to his wisdom and knowledge, you respect him and hesitant to talk him freely. So while asking a question from him, you break your sentences into small chunks and ask questions slowly. Chat GPT is that old man.

Simulate Persona

Suppose you ask a question like what degree should I Pursue, what career should I choose. Even What should do in my life?

Without context and with poor prompt ,it would be difficult to answer and chatGPT would fail. If we prompt our question like:

Where I live?

What are my interests?

What degree I have qualified?

This is my financial positon?

And so on-------------

After knowing these questions, I would have a context in my mind to suggest you a career choice.

In case of abdominal pain, instead of directly telling the doctor that I have pain in my abdomen, if you elaborate about your age, previous history etc it would be easy for the doctor to deal with your abdominal pain.

So the first prompt is to create the Context (Who Am I)  that is called Simulate Persona.

Task

Step to complete the Prompt

Constraints

Goal

Format Output

Number of Paragraphs, Image or text.



An easy approach is that when you are raising a query from Chat GPT , Imagine you are talking with a human. Giving the more than required or less than required information will create the confusion.

If you learn the art of selection of words and sequence of words , you will be successful in every field of life. Same is the case with Chat GPT if you select the proper words with proper sequence you will consume less tokens.

You pay after accumulating the token of input and output.

PLAYGROUND

Playground AI is a web app that allows users to create AI art and share it with the community. It offers free Dall-E2 image generation, automated art style prompts, free image upscaling, image saving on the cloud, and social media for AI-generated images.

Let us explain it with an example. You instruct your child, Servant or junior etc that be careful while conversation with me. We often correct him and advise him not repeat it again.

Play ground is free.







Frequency penalty works by lowering the chances of a word being selected again the more times that word has already been used. Presence penalty does not consider how frequently a word has been used, but just if the word exists in the text

The OpenAI Frequency Penalty setting is used to adjust how much frequency of tokens in the source material will influence the output of the model.

Frequency_penalty and presence_penalty are two parameters that can be used when generating text with language models, such as GPT-3.

·        Frequency_penalty: This parameter is used to discourage the model from repeating the same words or phrases too frequently within the generated text. It is a value that is added to the log-probability of a token each time it occurs in the generated text. A higher frequency_penalty value will result in the model being more conservative in its use of repeated tokens.

·        Presence_penalty: This parameter is used to encourage the model to include a diverse range of tokens in the generated text. It is a value that is subtracted from the log-probability of a token each time it is generated. A higher presence_penalty value will result in the model being more likely to generate tokens that have not yet been included in the generated text.

Both of these parameters can be adjusted to influence the overall quality and diversity of the generated text. The optimal values for these parameters may vary depending on the specific use case and desired output.

Temperature controls randomness, so a low temperature is less random (deterministic), while a high temperature is more random.

More technically, a low temperature makes the model more confident in its top choices, while temperatures greater than 1 decrease confidence in its top choices. An even higher temperature corresponds to more uniform sampling (total randomness). A temperature of 0 is equivalent to argmax/max likelihood, or the highest probability token.

top_p computes the cumulative probability distribution, and cut off as soon as that distribution exceeds the value of top_p. For example, a top_p of 0.3 means that only the tokens comprising the top 30% probability mass are considered.


Comments

Post a Comment

Popular posts from this blog

Topic-18 | Evaluation Metrics for different model

Image
  Lecture 18 | Evaluation Metrics for different model   Our previous topic was Linear Regression. Linear Regression is used to predict a number from continuous nominal or numeric values. The number we predict can be any number, it can be a positive number or a negative number, it can be a high value or a low value.   A graphical representation of linear regression was also shown.   In graphical representation there is plot, a studio plane, data points and a line that must have most of the data points around it. In order to measure the performance of graph, we also discussed the Mean Squared error and in order to nullify the effect of resultant zero root mean square error. In colab we trained a modal over a data set. An issue raised during its execution that the resultant value of Mean Squared Error was very high. Our today’s topic will start from the finding the reasons of this high value of Mean Squared Error. ...
Read more

Topic 22 | Neural Networks |

Image
  AI Free Basic Course | Lecture 22 | Neural Networks  In today lecture we will continue the Neural Network; in last lecture we gave the basic idea about what is to expect from neural network. Today we will explain the hyper parameters and its components. We will see the affect of the hyper parameter on the neural network. We will use playground and practice on notebook. At the end of lecture, we will implement a basic network using Tensor Flow. Today we will go through the things and would discuss the notebook in detail in tomorrow’s lecture that what does construct means in dense of Tensor flow, what does the layers do and what is loss function. What are possibilities to add in hidden layer and output layer. After practicing in TensorFlow you will get better idea about things. So, we will implement a neural network in TensorFlow and run it to see the outcome. Suppose we have few balls and our target is to through them in a basket placed at a distance of 10 meter. When ...
Read more

Topic 17 | Linear regression using Sklearn

Image
    AI Free Basic Course | Lecture 17 | Linear regression using Sklearn   Regression is the technique of supervised learning. There are two techniques in supervised learning, one is classification and other is regression. In classification we predict category / nominal value. Category or nominal value means something that has label; for example smoker /nonsmoker, cat /dog etc and we have to convert them into numerical value by effort. In prediction, we use the word positive and negative. For example, if a picture is given to the model and how it will predict whether it is picture of cat or not. Positive value in this would be that it is cat as we want to recognize cat. In medical test, the result is either positive or negative. So positive value is the value in which we are more interested to predict. Remember it does not depend upon  whether covid report is positive nor not, it depends what we have set the algorithm about which we are intereste...
Read more