Commit
•
b8ede3c
1
Parent(s):
4afe202
Update README.md
Browse files
README.md
CHANGED
@@ -26,7 +26,25 @@ base_model: google/gemma-7b
|
|
26 |
- **Finetuned from model:** [google/gemma-7b](https://huggingface.co/google/gemma-7b)
|
27 |
- **Contact**: For questions and comments about the model, please email `karakuri-rd@karakuri.ai`
|
28 |
|
29 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
|
31 |
```python
|
32 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
|
26 |
- **Finetuned from model:** [google/gemma-7b](https://huggingface.co/google/gemma-7b)
|
27 |
- **Contact**: For questions and comments about the model, please email `karakuri-rd@karakuri.ai`
|
28 |
|
29 |
+
## Usage
|
30 |
+
|
31 |
+
KARAKURI LM 7B APM v0.1 is a attribute prediction model that rates model responses on various aspects that makes a response desirable.
|
32 |
+
|
33 |
+
Given a conversation with multiple turns between user and assistant, the model rates the following attributes (between 0 and 4) for every assistant turn.
|
34 |
+
|
35 |
+
- helpfulness: Overall helpfulness of the response to the prompt.
|
36 |
+
- correctness: Inclusion of all pertinent facts without errors.
|
37 |
+
- coherence: Consistency and clarity of expression.
|
38 |
+
- complexity: Intellectual depth required to write response (i.e. whether the response can be written by anyone with basic language competency or requires deep domain expertise).
|
39 |
+
- verbosity: Amount of detail included in the response, relative to what is asked for in the prompt.
|
40 |
+
- quality: Perceived goodness of response
|
41 |
+
- toxicity: Undesirable elements such as vulgar, harmful or potentially biased response
|
42 |
+
- humor: Sense of humor within response
|
43 |
+
- creativity: Willingness to generate non-conventional response
|
44 |
+
|
45 |
+
The first five are derived from HelpSteer, while the remaining four are derived from OASST2.
|
46 |
+
|
47 |
+
You can run the model using the 🤗 Transformers:
|
48 |
|
49 |
```python
|
50 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|