Update README.md
Browse files
README.md
CHANGED
@@ -5,6 +5,29 @@ This model is a fine-tuned version of [microsoft/phi-3-mini-4k-instruct](https:/
|
|
5 |
|
6 |
## Example Conversations
|
7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
### Example 1: Astrid Character
|
9 |
Note some portions of conversation redacted for public README purposes
|
10 |
|
|
|
5 |
|
6 |
## Example Conversations
|
7 |
|
8 |
+
### Prompt Template Guidelines
|
9 |
+
|
10 |
+
**Important: This model requires a specific prompt format to function properly.**
|
11 |
+
|
12 |
+
For best results, follow this exact template structure:
|
13 |
+
|
14 |
+
```
|
15 |
+
Character's Persona: [detailed persona description]
|
16 |
+
User: [user message]
|
17 |
+
Assistant:
|
18 |
+
```
|
19 |
+
|
20 |
+
The model will complete the response after "Assistant:". This specific format is critical - using standard chat templates or different formatting will result in poor or nonsensical outputs.
|
21 |
+
|
22 |
+
### Key Points:
|
23 |
+
- Start with the character description using "Character's Persona:" format
|
24 |
+
- Each user message must be prefixed with "User: "
|
25 |
+
- Each model response will be after "Assistant:"
|
26 |
+
- The model often includes emoji prefixes (👅, 😈, 🔥, etc.) in its responses
|
27 |
+
- Keep the exact formatting including newlines as shown above
|
28 |
+
|
29 |
+
Detailed code examples showing implementation are provided in the Usage section below.
|
30 |
+
|
31 |
### Example 1: Astrid Character
|
32 |
Note some portions of conversation redacted for public README purposes
|
33 |
|