Datasetpublic

TurquoiseKitty/ultrafeedback-23612-Gemma-27b-scoring

Ultrafeedback 23612 Gemma 27B Scoring Languages The dataset is in English. Dataset Structure Data Instances A typical example looks like this: { "index_col": 0, "prompt": "Teacher: In this task, you are given a word. You should respond with a valid sentence which contains the given word. Make sure that the sentence is grammatically correct. You may use the word in a different tense than is given. For example, you may use the word… See the full description on the dataset page: https://huggingface.co/datasets/TurquoiseKitty/ultrafeedback-23612-Gemma-27b-scoring.

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes16downloads

Dataset Card

Ultrafeedback 23612 Gemma 27B Scoring

Languages

The dataset is in English.

Dataset Structure

Data Instances

A typical example looks like this:

json

{
  "index_col": 0,
  "prompt": "Teacher: In this task, you are given a word. You should respond with a valid sentence which contains the given word. Make sure that the sentence is grammatically correct. You may use the word in a different tense than is given. For example, you may use the word 'ended' in the output where the given input word is 'end'.\nTeacher: Now, understand the problem? If you are still confused, see the following example:\ndust\nSolution: Another one bites the dust.\nReason: The output sentence is grammatically correct and contains the given word. Hence, it is a valid response to the question.\n\nNow, solve this instance: loaf\nStudent:",
  "chosen": "After a long day at work, Mary decided to treat herself to a warm, freshly baked loaf of bread.",
  "rejected": "I learned how to bake a loaf of bread in baking class.",
  "chosen_score": -2.390625,
  "rejected_score": -3.390625
}

Data Fields

—index_col: int64
—prompt: string
—chosen: string
—rejected: string
—chosen_score: float64
—rejected_score: float64

Data Splits

Split	Examples
train	23,612

Dataset Creation

Source Data

This dataset was manually annotated and prepared for machine learning tasks.

Data Collection and Processing

The data has been processed and stored in Arrow format for efficient loading and processing.

Usage

Loading the Dataset

python

from datasets import load_dataset

# Load the entire dataset
dataset = load_dataset("TurquoiseKitty/ultrafeedback-23612-Gemma-27b-scoring")

# Load specific split
train_data = load_dataset("TurquoiseKitty/ultrafeedback-23612-Gemma-27b-scoring", split="train")

Contributions

Dataset created and annotated by TurquoiseKitty.