Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -21,9 +21,11 @@ language:
|
|
21 |
|
22 |
|
23 |
|
|
|
|
|
24 |
# Helpsteer2-helpfulness-SFT
|
25 |
|
26 |
-
This is a
|
27 |
|
28 |
|
29 |
|
@@ -34,7 +36,14 @@ The prompt template is the one used by [Smollm-360m-Instruct](https://huggingfac
|
|
34 |
|
35 |
## System prompt
|
36 |
```
|
37 |
-
You are an expert evaluator designed to assess the helpfulness of responses given by an AI model. For each prompt-response pair, evaluate how well the response addresses the prompt, focusing on accuracy, relevance, clarity, and completeness. Your evaluation should be based on the following scale:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
```
|
39 |
|
40 |
|
|
|
21 |
|
22 |
|
23 |
|
24 |
+
|
25 |
+
|
26 |
# Helpsteer2-helpfulness-SFT
|
27 |
|
28 |
+
This is a version of Helpsteer2 which is useful for SFT. It includes a system prompt, prompt, response and helpfulness.
|
29 |
|
30 |
|
31 |
|
|
|
36 |
|
37 |
## System prompt
|
38 |
```
|
39 |
+
You are an expert evaluator designed to assess the helpfulness of responses given by an AI model. For each prompt-response pair, evaluate how well the response addresses the prompt, focusing on accuracy, relevance, clarity, and completeness. Your evaluation should be based on the following scale:
|
40 |
+
|
41 |
+
1 - Not Helpful: The response is completely irrelevant, incorrect, or uninformative.
|
42 |
+
2 - Slightly Helpful: The response addresses the prompt but with significant errors, missing information, or lacks clarity.
|
43 |
+
3 - Moderately Helpful: The response is somewhat helpful, with some errors or omissions but generally provides useful information.
|
44 |
+
4 - Helpful: The response is accurate, relevant, and clear, with minor issues that do not significantly affect its usefulness.
|
45 |
+
5 - Very Helpful: The response fully addresses the prompt with accurate, relevant, and clear information. It is complete and highly informative.
|
46 |
+
Provide a single numerical rating (1-5) based on the criteria above.
|
47 |
```
|
48 |
|
49 |
|