nefro313/hr_question_llama-2-formatted

HR Questions and Answers Dataset for LLaMA 2

Overview

This dataset contains 26 carefully crafted HR-related questions along with corresponding answers formatted specifically for fine-tuning large language models like LLaMA 2. The questions cover a wide range of HR scenarios, such as interview preparation, workplace ethics, and professional skills. Each data point has been preprocessed into a format suitable for instruction-tuned models, making it ideal for tasks like text generation and conversational AI.

Dataset Structure

Format

The dataset has been processed into a JSONL (JSON Lines) format and uploaded to Hugging Face in Parquet format for efficient storage and access. Each data instance is structured as follows:

• Instruction: The question or prompt (e.g., "Tell me about yourself.")
• Response: A well-articulated answer to the question, providing valuable insights or guidance.

Example:

{
    "instruction": "Tell me about yourself.",
    "response": "This is often the first question to break the ice. Candidates should provide a concise summary of their professional background, skills, and achievements. For example: 'I have over 5 years of experience in software development, specializing in backend systems...'
}

Data Splits

• Train Dataset: Contains 80% of the questions and answers.
• Test Dataset: Contains 20% of the questions and answers for evaluation purposes.

Usage

Download the Dataset

You can download the dataset directly from its Hugging Face repository:

from datasets import load_dataset

dataset = load_dataset("nefro313/hr_question_llama-2-formatted")

Fine-tuning

This dataset is suitable for fine-tuning instruction-tuned language models. Below is an example of how to fine-tune using Hugging Face's transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer, TrainingArguments, Trainer
from datasets import load_dataset

# Load dataset
dataset = load_dataset("nefro313/hr_question_llama-2-formatted")

# Load model and tokenizer
model_name = "<your-llama-2-model>"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Preprocess data for training
def preprocess_function(examples):
    return tokenizer(examples["instruction"] + examples["response"], truncation=True)

tokenized_datasets = dataset.map(preprocess_function, batched=True)

# Define training arguments
training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    learning_rate=2e-5,
    per_device_train_batch_size=4,
    num_train_epochs=3,
    weight_decay=0.01,
    save_strategy="epoch",
    logging_dir="./logs",
    push_to_hub=True
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["test"]
)

trainer.train()

Citation

If you use this dataset, please cite it as follows:

@dataset{nefro313_hr_dataset,
  author = {Robin K. Philip},
  title = {HR Questions and Answers Dataset for LLaMA 2},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/nefro313/hr_question_llama-2-formatted}
}

License

This dataset is licensed under the Apache 2.0 License. You are free to use it for both academic and commercial purposes.

Contributions

Contributions are welcome! If you have additional HR-related questions or answers that can improve this dataset, feel free to submit a pull request or reach out via the Hugging Face repository.