Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,15 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
# Date-DPO-v1
|
| 6 |
+
|
| 7 |
+
DPO dataset aiming to reduce output verbosity and "GPT-speak."
|
| 8 |
+
|
| 9 |
+
## Method
|
| 10 |
+
|
| 11 |
+
Categories and prompts spanning various topics were selected and generated.
|
| 12 |
+
|
| 13 |
+
ChatGPT 3.5's one-shot answers were selected as the `rejected` responses.
|
| 14 |
+
|
| 15 |
+
flammen19X-mistral-7B was used to generate `chosen` prompts. Many of these responses were prompted with "respond conversationally/succinctly" and modified to be shorter.
|