metadata
dataset_info:
features:
- name: task_id
dtype: string
- name: prompt
dtype: string
- name: canonical_solution
dtype: string
- name: test
dtype: string
- name: entry_point
dtype: string
- name: difficulty_scale
dtype: string
splits:
- name: test
num_bytes: 178475
num_examples: 151
download_size: 72166
dataset_size: 178475
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
qiskit_humaneval
qiskit_humaneval is a dataset for evaluating LLM's at writing Qiskit code.
Dataset
The Qiskit HumanEval dataset is inspired by OpenAI's HumanEval. It includes 151 Qiskit problems written by human Qiskit advocates.
Terms of use
- Terms of use: https://quantum.ibm.com/terms
- Privacy policy: https://quantum.ibm.com/terms/privacy