Add 1 files
Browse files
README.md
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
annotations_creators: []
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
language_creators: []
|
| 6 |
+
license: []
|
| 7 |
+
multilinguality:
|
| 8 |
+
- monolingual
|
| 9 |
+
pretty_name: 'adaption-mixed_reasoning_qa'
|
| 10 |
+
size_categories:
|
| 11 |
+
- n<1K
|
| 12 |
+
source_datasets: []
|
| 13 |
+
tags:
|
| 14 |
+
- adaption
|
| 15 |
+
- preference-training
|
| 16 |
+
- math
|
| 17 |
+
- how-to
|
| 18 |
+
task_categories: []
|
| 19 |
+
task_ids: []
|
| 20 |
+
---
|
| 21 |
+
|
| 22 |
+

|
| 23 |
+
|
| 24 |
+
This dataset is a remastered version prepared using [Adaption's](https://frontend-pharos-app-dev.vercel.app/app/auth) Adaptive Data platform.
|
| 25 |
+
|
| 26 |
+
# adaption-mixed_reasoning_qa
|
| 27 |
+
|
| 28 |
+
This dataset contains a diverse collection of prompt-completion pairs covering mathematical problem-solving, logical reasoning, and general knowledge queries. The content ranges from algebraic derivations and number theory proofs to practical survival strategies and basic sorting tasks. Responses vary in length and detail, including step-by-step explanations, concise answers, and structured lists.
|
| 29 |
+
|
| 30 |
+
### Dataset size
|
| 31 |
+
|
| 32 |
+
There are 990 data points in this dataset. This is a preference training dataset.
|
| 33 |
+
|
| 34 |
+
### Quality of Remastered Dataset
|
| 35 |
+
|
| 36 |
+
The final quality is A, with a relative quality improvement of 53.3%.
|
| 37 |
+
|
| 38 |
+
### Domain
|
| 39 |
+
- Math (54%)
|
| 40 |
+
- How-to (46%)
|
| 41 |
+
|
| 42 |
+
### Language
|
| 43 |
+
- English (100%)
|
| 44 |
+
|
| 45 |
+
### Tone
|
| 46 |
+
- Clear (24%)
|
| 47 |
+
- Practical (22%)
|
| 48 |
+
- Educational (18%)
|
| 49 |
+
|
| 50 |
+
### Evaluation Results
|
| 51 |
+
|
| 52 |
+
- **Quality Gains:**
|
| 53 |
+
<img src="https://proteus-dev-public.s3.us-east-1.amazonaws.com/temp/e2eee0a7-dfd8-48d2-b0fd-d15e687be8a5.png" alt="QualityGains" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|
| 54 |
+
|
| 55 |
+
- **Grade Improvement:**
|
| 56 |
+
<img src="https://proteus-dev-public.s3.us-east-1.amazonaws.com/temp/19d55443-dbed-46cb-b849-416c232a7b78.png" alt="Grade" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|
| 57 |
+
|
| 58 |
+
- **Percentile Chart:**
|
| 59 |
+
<img src="https://proteus-dev-public.s3.us-east-1.amazonaws.com/temp/c6a67f6c-9b11-4f08-ad78-253fe3c27fc4.png" alt="Percentile Chart" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />
|