Please enable JavaScript.
Coggle requires JavaScript to display documents.
GH Copilot Exam - Coggle Diagram
GH Copilot Exam
Responsible AI
-
Using AI responsibly is crucial to ensure that its impact is positive and aligned with ethical standards
Quality Assurance: Users need to review and test generated code, because training data is also based on public repositories
-
Code Licensing: Developers must ensure that any code they use complies with relevant licenses and does not infringe on others' intellectual property [TO VERIFY FILTER]
Documentation
-
-
Copilot Chat
- Chat view: workspace Explain this project
- Chat view: workspace /explain Explain the dependencies of this project
- Chat view: workspace /explain #file:program.cs Explain how this file is used in the project
- Inline chat: /explain #selection Explain how this method works
- Inline chat: /explain Explain this code block
-
-
Prompt
Strategies
Start general, then get specific
-
-
-
-
-
-
-
-
Agents
terminal agent helps you chat with GitHub Copilot to interact with the terminal
workspace which is aware of your entire workspace, allowing you to ask questions about the entire project
Although you can be specific with workspace, by default GitHub Copilot uses open files in your text editor as additional context.
-
Handles Data
LLMs
-
-
-
-
Fine Tuning
LoRA fine tuning
-
The original model remains the same, which saves time and resources.
-
-
It involves training the model on a smaller, task-specific dataset, known as the target dataset, while using the knowledge and parameters gained from a large pretrained dataset, referred to as the source model.
- Secure prompt transmission and context gathering
-
-
-
-
-
- Repeat for subsequent prompts
-
Privacy
Secure transmission/encryption: including both contextual data about the code and file being edited (“prompts”) and data about the user’s actions (“user engagement data”). The transmitted data is encrypted both in transit and at res
-