detail prompt format
Browse files
README.md
CHANGED
@@ -18,6 +18,15 @@ This is a linear merge with the following composition:
|
|
18 |
- [NeverSleep/MiquMaid-v3-70B](https://huggingface.co/NeverSleep/MiquMaid-v3-70B) at weight 0.3
|
19 |
- [maywell/miqu-evil-dpo](https://huggingface.co/maywell/miqu-evil-dpo) at weight 0.2
|
20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
## Merge Configuration
|
22 |
|
23 |
The following YAML configuration was used to produce this model:
|
|
|
18 |
- [NeverSleep/MiquMaid-v3-70B](https://huggingface.co/NeverSleep/MiquMaid-v3-70B) at weight 0.3
|
19 |
- [maywell/miqu-evil-dpo](https://huggingface.co/maywell/miqu-evil-dpo) at weight 0.2
|
20 |
|
21 |
+
## Prompt format
|
22 |
+
The model responds well to general-purpose prompt formats such as Alpaca. Alternatively, I suggest trying the following format:
|
23 |
+
|
24 |
+
```
|
25 |
+
<message from='system'>{your system prompt here}</message><message from='user'>{user prompt here}</message><message from='bot'>{bot response here}</message><message from='user'>{user prompt here}</message><message from='bot'>{bot response here}</message> [... and so on ...]
|
26 |
+
```
|
27 |
+
|
28 |
+
This format is readily understood by the model, and leads to the expected high-quality responses. Note the lack of newlines `\n` - they are not necessary and might actually make it harder for the model to follow along.
|
29 |
+
|
30 |
## Merge Configuration
|
31 |
|
32 |
The following YAML configuration was used to produce this model:
|