Beginner’s guide to using ChatGPT

There’s lots of talk about AI and how it can be used – and that includes politicians and Bob down the pub! It’s definitely no longer restricted to academia and IT conferences. Earlier this year, it seemed like everyone was using ChatGPT to see what it can do, and, of course, there have been worries that students would use it to write essays, and studios would use it to write scripts for TV shows. Having said that, more recently, I have found quite a few people who have been keeping up with the conversation, but haven’t actually had a go yet. So, for them, I thought I’d put together this beginner’s guide to using ChatGPT.

ChatGPT is an advanced natural language processing model developed by OpenAI. It is designed to generate human-like responses to prompts or questions. Whether you’re a beginner or an experienced user, this guide will help you navigate the ChatGPT interface and make the most out of your conversational interactions. Here’s a step-by-step approach to get you started.

Step 1: Accessing ChatGPT

To access ChatGPT, visit the OpenAI website (https://www.openai.com) and locate the “GPT-3 Playground”. Once you’re on the playground, you’ll see a text box where you can interact with the model.

Step 2: Prompt generation

Begin by entering a prompt in the text box to start a conversation. The system will generate a response based on the context and information provided. The prompt can be a question, a sentence, or a statement that guides the conversation.

For example, you can start with a simple prompt like: “Tell me a joke”. ChatGPT will then generate a response with a joke. You can build on this conversation by asking further questions or statements related to the generated response.

Step 3: Iterative conversation

One key feature of ChatGPT is its ability to maintain continuity across multiple messages. To achieve this, precede your input with a user or system name followed by a colon (“:”). For example, “User: What’s your favourite colour?” or “Assistant: How tall are you?”

By maintaining the user and assistant distinction, the model can better understand and respond to the conversation’s context. Iterative conversation allows you to engage in extended interactions and improves the model’s understanding of the intended conversation flow.

Step 4: System-level instructions

To guide the model’s behaviour, you can provide high-level instructions. These instructions help shape the response based on the desired outcome. For example, you can specify, “Assistant: Answer the question as if you’re a Shakespearean character”.

By giving system instructions, you can explore creative responses from the model. Experiment with these instructions to see how they impact the flavour or style of the generated text.

Step 5: Managing response length

You can control the length of the generated response using the ‘Max tokens’ setting. The more tokens you allow, the longer and more detailed the response will be. However, increasing the token limit also affects the model’s response time. Adjust this setting based on your preference and the desired level of detail.

Step 6: Learning from examples

Training the model to generate precise responses requires providing it with clear instructions. If a response is not satisfactory, you can use examples to illustrate how you want it to be improved.

For instance, you can mention previous correct or desired responses during conversation. This helps ChatGPT understand the desired output, leading to better performance and more accurate replies.

Step 7: Experimentation and refinement

Using ChatGPT is an iterative process. Experimentation and refining the prompts or instructions will improve the quality of generated responses. You can try different prompts, lengths, system instructions, or even restructuring your conversation to achieve better outcomes.

Remember, refining prompts and iterating through conversations will help you discover the best approaches to obtain the desired results.

Step 8: Ethical use and monitoring

While ChatGPT offers exciting possibilities, it is crucial to use it responsibly. OpenAI encourages users to follow their usage guidelines to ensure ethical and appropriate use of the model. Stay vigilant and monitor the system-generated content to prevent any misuse or biased outcomes.

Conclusion

ChatGPT is a powerful tool for generating human-like responses to prompts and holding natural conversations. By following this beginner’s guide, you can get started with ease and explore the capabilities of this remarkable language model. Remember to experiment, iterate, and use it ethically as you delve into the world of conversational AI.

You’ll not be surprised to learn that I asked ChatGPT to write the article. In the past, I have asked it to write a sonnet about mainframes, which I posted on LinkedIn, and which PJ (Pasquale) Catalano, IBM Z & LinuxONE Test Architect, recorded and also posted on LinkedIn.

What would I have added? I may have stressed that prompts really need to be specific to get a good response. You can ask ChatGPT to act in a particular role (eg customer support or HR manager), and you can ask it to perform a task (eg writing a job description of responding to a complaint). You can specify the format (eg bullet points or a flow chart), and you can get it to use a particular tone (eg engaging or serious). Also, you need to specify the objective.

ChatGPT integrates with business tools, it’s useful for data analysis and computations, and it can handle large numbers of queries at the same time, making it scalable. If you haven’t used it yet, then definitely take a look and see what all the hype has been about.

It’s worth noting that OpenAI recently launched ChatGPT Enterprise, which is meant to be a more secure version because it can be compliant with corporate privacy regulation, ie it supports features like SSO and usage insights.