dataformer/dolly-llama-qa

---
language:
- en
size_categories:
- 1K<n<10K
language_bcp47:
- en-US
configs:
- config_name: llama_3.1_8B
data_files:
- split: train
path: llama_3-1_8B_Instruct.jsonl
- config_name: llama_3_8B
data_files:
- split: train
path: llama_3_8B_Instruct.jsonl
task_categories:
- text-generation
- question-answering
tags:
- synthetic
---
# Dataset Card for dolly-llama-qa

<!-- Provide a quick summary of the dataset. -->

This dataset has been created with [dataformer](https://github.com/DataformerAI/dataformer).

## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->

The dolly-llama-qa dataset is a synthetic QA pair dataset created using the context from [databricks-dolly-15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k). We used [Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) and [Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) models for the generation and evolution part. Openai's gpt-4o was used for evaluating the refined questions and refined answers.

## Dataset Columns

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->

* `context`: the context taken from [databricks-dolly-15k](https://huggingface.co/datasets/databricks/databricks-dolly-15k)
* `seed_question`: the seed question generated from the context using the instruct model.
* `refined_question`: the question after evolving the seed question using the instruct model.
* `initial_answer`: the answer to the refined_question generated using the instruct model
* `refined_answer`: the answer after evolution of the initial answer using the instruct model.
* `question_quality`: the quality of the question on a scale of 1-10 evaluated using gpt-4o
* `answer_quality`: the quality of the answer on a scale of 1-10 evaluated using gpt-4o