Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,23 @@
|
|
| 1 |
---
|
| 2 |
license: cc-by-4.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: cc-by-4.0
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
This model judges if a given output is sufficient to recreate a given instruction.
|
| 6 |
+
|
| 7 |
+
It's useful for filtering data to train a reverse instruct model. It could also have applications around determining if an output/instruction pair is linked, or around quality filtering data (data where the instruction can be recreated from the output might be higher quality).
|
| 8 |
+
|
| 9 |
+
The model is a binary classifier trained on top of Python 160m with 50k examples for 1 epoch. The final validation loss is TBD.
|
| 10 |
+
|
| 11 |
+
To use it, pass in this prompt format:
|
| 12 |
+
|
| 13 |
+
```
|
| 14 |
+
Output
|
| 15 |
+
|
| 16 |
+
{output}
|
| 17 |
+
|
| 18 |
+
Instruction
|
| 19 |
+
|
| 20 |
+
{instruction}
|
| 21 |
+
```
|
| 22 |
+
|
| 23 |
+
Output should be the output from a model, and instruction should be the instruction that generated the output. The model will return a 0-1 score indicating how effectively the instruction can be recreated.
|