[task] Personalized Dialog

### 🚀 The Task

Given a "role" tag, the model should be able to adapt its "personality" to user liking in order to promote more productive pair programming interactions.

Concerns: This could lead to malicious use of the model.

### Example

```
Role: You are a comical but highly proficient software engineer.

Can you help me with the following issues?
```

### Additional Notes

* Reference: [personaGPT](https://github.com/illidanlab/personaGPT) and [character.ai](https://beta.character.ai/)
* Google Sheets Reference: [CarperAI: Instruct-GPT Tasks](https://docs.google.com/spreadsheets/d/1q6I0PLcgvNjcls3z7diUOoDGmNojUW-m6KhX2ijuSEE/edit?usp=sharing)