---
dataset_info:
  features:
  - name: code
    dtype: string
  - name: level
    dtype: int64
  - name: name
    dtype: string
  - name: problem_id
    dtype: int64
  splits:
  - name: level_1
    num_bytes: 133215
    num_examples: 100
  - name: level_2
    num_bytes: 114519
    num_examples: 100
  - name: level_3
    num_bytes: 177082
    num_examples: 50
  - name: level_4
    num_bytes: 15591
    num_examples: 20
  download_size: 115926
  dataset_size: 440407
configs:
- config_name: default
  data_files:
  - split: level_1
    path: data/level_1-*
  - split: level_2
    path: data/level_2-*
  - split: level_3
    path: data/level_3-*
  - split: level_4
    path: data/level_4-*
pretty_name: KernelBench
---
# Dataset Card for Dataset Name

<!-- Provide a quick summary of the dataset. -->

This is the copy from Stanford's KernelBench (https://huggingface.co/datasets/ScalingIntelligence/KernelBench). 

## Dataset Details

- Level 1: 100 Problems
- Level 2: 100 Problems
- Level 3: 50 Problems
- Level 4: 20 Problems



## Plan:
1. We want to try and tackle the dataset as well at MBZUAI / Imperial College London.