issai/GPQA_Kazakh_Russian

---
dataset_info:
- config_name: kazakh
features:
- name: question
dtype: string
- name: choices
list: string
- name: answer
dtype: int64
- name: subdomain
dtype: string
splits:
- name: test
num_bytes: 1274749
num_examples: 1192
download_size: 550184
dataset_size: 1274749
- config_name: russian
features:
- name: question
dtype: string
- name: choices
list: string
- name: answer
dtype: int64
- name: subdomain
dtype: string
splits:
- name: test
num_bytes: 1319241
num_examples: 1192
download_size: 586919
dataset_size: 1319241
configs:
- config_name: kazakh
data_files:
- split: test
path: kazakh/test-*
- config_name: russian
data_files:
- split: test
path: russian/test-*
task_categories:
- question-answering
language:
- kk
size_categories:
- 1K<n<10K
license: cc-by-4.0
---

## Dataset Summary

These are the machine-translated Kazakh and Russian versions of the [GPQA (Graduate-Level Google-Proof Q&A Benchmark)](https://huggingface.co/datasets/Idavidrein/gpqa) dataset.

These datasets are used to test the world knowledge and problem-solving capabilities of large language models across a vast range of subjects in the Kazakh and Russian languages. Unlike general knowledge benchmarks, GPQA consists of extremely challenging science questions (biology, physics, and chemistry) written by experts. These questions are designed to be "Google-proof," meaning they are difficult for non-experts to answer even with unrestricted internet access, making this a rigorous test of a model's advanced reasoning and specialized scientific knowledge in Kazakh or Russian.