datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/tau/commonsense_qa.commonsense_qa_2.0https://github.com/allenai/csqa2
@article{talmor2022commonsenseqa,
title={CommonsenseQA 2.0: Exposing the limits of AI through gamification},
author={Talmor, Alon and Yoran, Ori and Bras, Ronan Le and Bhagavatula, Chandra and Goldberg, Yoav and Choi, Yejin and Berant, Jonathan},
journal={arXiv preprint arXiv:2201.05320},
year={2022}
}
CommonsenseQA-Explained-by-ChatGPTThis is a dataset with explanations from ChatGPT for the correct and incorrect answers in CommonsenseQA. The explanations are generated by prompting ChatGPT with answer keys and in-context examples. We expect this dataset to be an useful source for understanding the commonsense reasoning ability of LLMs or training other LMs.
CommonsenseQA1000COTCRAFT-CommonSenseQA
CRAFT-CommonSenseQA
This is a synthetic dataset generated with the CRAFT framework proposed in the paper CRAFT Your Dataset: Task-Specific Synthetic Data Generation Through Corpus Retrieval and Augmentation.
The correctness of the data has not been verified in detail, but training on this data and evaluating on human-curated commonsense question-answering data proved highly beneficial.
4 synthetic dataset sizes (S, M, L, XL) are available, and training on them yields consistent… See the full description on the dataset page: https://huggingface.co/datasets/ingoziegler/CRAFT-CommonSenseQA.commonsense_filtered
Dataset Summary
The commonsense reasoning tasks consist of 8 subtasks, each with predefined training and testing sets, as described by LLM-Adapters (Hu et al., 2023). The following table lists the details of each sub-dataset.
Train
Test
Information
BoolQ (Clark et al., 2019)
9427
3270
Question-answering dataset for yes/no questions
PIQA (Bisk et al., 2020)
16113
1838
Questions with two solutions requiring physical commonsense to answer
SIQA (Sap et al., 2019)
33410… See the full description on the dataset page: https://huggingface.co/datasets/fxmeng/commonsense_filtered.commonsense_cot_partial_raw
Commonsense QA CoT (Partial, Raw, No Human Annotation)
Dataset Summary
Seeded by the CommonsenseQA dataset (tau/commonsense_qa) this preliminary set randomly samples 1,000 question-answer
entries and uses Mixtral (mistralai/Mixtral-8x7B-Instruct-v0.1) to generate 3 unique CoT (Chain-of-Thought) rationales.
This was created as the preliminary step towards fine-tuning a LM (language model) to specialize on commonsense reasoning.
The working hypothesis, inspired by the… See the full description on the dataset page: https://huggingface.co/datasets/peterkchung/commonsense_cot_partial_raw.chatgpt4-commonsense-qa
Synthetic CommonSense
Generated using ChatGPT4, originally from https://huggingface.co/datasets/commonsense_qa
Notebook at https://github.com/mesolitica/malaysian-dataset/tree/master/question-answer/chatgpt4-commonsense
synthetic-commonsense.jsonl, 36332 rows, 7.34 MB.
Example data
{'question': '1. Seseorang yang bersara mungkin perlu kembali bekerja jika mereka apa?\n A. mempunyai hutang\n B. mencari pendapatan\n C. meninggalkan pekerjaan\n D. memerlukan… See the full description on the dataset page: https://huggingface.co/datasets/mesolitica/chatgpt4-commonsense-qa.commonsenseqa-bn
Dataset Summary
This is the Bangla translated version of the CommonsenseQA dataset. The dataset was translated using a new method called Expressive Semantic Translation (EST). This method combines both Google Machine Translation and LLM-based rewriting of the translation to enhance the expressiveness and semantic accuracy of the translated content.
Dataset Structure
Data instances
Defaults
An example of a 'train' looks as follows:
{… See the full description on the dataset page: https://huggingface.co/datasets/hishab/commonsenseqa-bn.commonsense_qa_zh
Commonsense QA Chinese Multiple-Choice Dataset
This dataset is a Chinese four-choice SFT version of tau/commonsense_qa. It is designed to supplement commonsense multiple-choice training data for benchmark tasks such as challenge_common_sense.
The original dataset is in English and contains five-choice commonsense questions. This release keeps only samples that can be aligned to the official four-choice benchmark format, translates the question and options into… See the full description on the dataset page: https://huggingface.co/datasets/choucsan/commonsense_qa_zh.commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/spacekat99/commonsense_qa.commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/psjadf/commonsense_qa.commonsense_qa-mt-pt
CommonsenseQA-PT
Portuguese machine translation of CommonsenseQA, a multiple-choice question answering dataset that requires commonsense reasoning.
Translated using a Finetuned GemmaX2-9B for pt-PT.
Original Dataset: https://huggingface.co/datasets/tau/commonsense_qa
Note: This dataset is machine translated and may contain translation errors or artifacts.
This dataset is provided as part of the AMALIA project and is included in AMALIA-Bench, a comprehensive benchmark… See the full description on the dataset page: https://huggingface.co/datasets/amalia-llm/commonsense_qa-mt-pt.Common-Sense-Reasoning
🇰🇿 Kazakh General Inquiry and FAQ Dataset
📖 Overview
This dataset contains 1,000 high-quality question-and-answer pairs in the Kazakh language. It is designed to train models on providing helpful, natural, and informative responses to common inquiries.
📊 Dataset Statistics
General Metrics
Metric
Count
Total Samples
1,000
Total Words (approx.)
74,013
Avg. Words per Sample
74
Word Count Distribution… See the full description on the dataset page: https://huggingface.co/datasets/farabi-lab/Common-Sense-Reasoning.commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/AdrianSword/commonsense_qa.commonsense_qa-ka
commonsense_qa-ka
Georgian translation of the CommonsenseQA benchmark.
Dataset Summary
Property
Value
Examples
1,221
Splits
validation
Languages
Georgian, English
Task
Commonsense Reasoning
Data Fields
id: Example ID
question: Question (English)
question_concept: Core concept
choices: Answer choices (English)
answerKey: Correct answer key
question_ka: Question (Georgian)
choices_ka: Answer choices (Georgian)
Translation… See the full description on the dataset page: https://huggingface.co/datasets/tbilisi-ai-lab/commonsense_qa-ka.aihub_mrc_commonsense
Dataset Card for "mrc_aihub_common_sense"
일반 상식
commonsense_qa-IDCommonsenseQA-ID is Indonesian translation version of CommonsenseQA, a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/JazonDeng/commonsense_qa.commonsense_cot_partial_raw
Commonsense QA CoT (Partial, Raw, No Human Annotation)
Dataset Summary
Seeded by the CommonsenseQA dataset (tau/commonsense_qa) this preliminary set randomly samples 1,000 question-answer
entries and uses Mixtral (mistralai/Mixtral-8x7B-Instruct-v0.1) to generate 3 unique CoT (Chain-of-Thought) rationales.
This was created as the preliminary step towards fine-tuning a LM (language model) to specialize on commonsense reasoning.
The working hypothesis, inspired by the… See the full description on the dataset page: https://huggingface.co/datasets/spacekat99/commonsense_cot_partial_raw.hads-physical-commonsense
HADS – Human Action and Decision Sense (حَدس)
📄 Paper: HADS: A Large-Scale Parallel Benchmark for Physical Commonsense Reasoning — IEEE Access, 2026 (doi:10.1109/ACCESS.2026.3705337)
HADS is a large-scale Arabic parallel adaptation of the English
PIQA benchmark for physical commonsense reasoning.
The name derives from the Arabic word حَدس (hads), meaning physical intuition
or gut sense — the tacit embodied knowledge the benchmark measures.
Dataset summary… See the full description on the dataset page: https://huggingface.co/datasets/IWAN/hads-physical-commonsense.commonsense_qa
Dataset Card for "commonsense_qa"
Dataset Summary
CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge
to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers.
The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation
split, and "Question token split", see paper for details.… See the full description on the dataset page: https://huggingface.co/datasets/chouxtteok/commonsense_qa.