README.md

---
license: cc-by-sa-4.0
dataset_info:
  features:
  - name: path
    dtype: string
  - name: max_stars_owner
    dtype: string
  - name: min_issues_main_language
    dtype: string
  - name: max_forks_main_language
    dtype: string
  - name: max_forks_commit_sha
    dtype: string
  - name: min_issues_repo_id
    dtype: float64
  - name: non_alpha_syms
    dtype: int64
  - name: min_stars_owner
    dtype: string
  - name: num_of_duplicates
    dtype: int64
  - name: max_stars_languages_distribution
    dtype: string
  - name: std_line_len
    dtype: float64
  - name: pii_info
    list:
    - name: end
      dtype: int64
    - name: entity_group
      dtype: string
    - name: score
      dtype: float64
    - name: start
      dtype: int64
    - name: word
      dtype: string
  - name: cluster_id
    dtype: int64
  - name: non_empty_len
    dtype: int64
  - name: min_stars_languages_distribution
    dtype: string
  - name: min_forks_name
    dtype: string
  - name: min_forks_repo_id
    dtype: int64
  - name: min_issues_owner
    dtype: string
  - name: max_issues_name
    dtype: string
  - name: owner
    dtype: string
  - name: extension
    dtype: string
  - name: min_issues_name
    dtype: string
  - name: min_forks
    dtype: int64
  - name: max_issues_main_language
    dtype: string
  - name: repo_id
    dtype: int64
  - name: license
    dtype: string
  - name: max_stars_path
    dtype: string
  - name: max_stars_commit_sha
    dtype: string
  - name: avg_line_len
    dtype: float64
  - name: max_stars_main_language
    dtype: string
  - name: min_stars
    dtype: int64
  - name: partition_0
    dtype: string
  - name: min_stars_repo_id
    dtype: int64
  - name: max_len
    dtype: int64
  - name: is_fork
    dtype: bool
  - name: max_stars_name
    dtype: string
  - name: max_forks
    dtype: int64
  - name: max_issues_languages_distribution
    dtype: string
  - name: min_stars_commit_sha
    dtype: string
  - name: total_lines
    dtype: int64
  - name: min_forks_commit_sha
    dtype: string
  - name: max_forks_name
    dtype: string
  - name: total_symbols
    dtype: int64
  - name: languages_distribution
    dtype: string
  - name: max_forks_repo_id
    dtype: int64
  - name: content
    dtype: string
  - name: issues
    dtype: float64
  - name: max_issues_commit_sha
    dtype: string
  - name: max_forks_path
    dtype: string
  - name: main_language
    dtype: string
  - name: forks
    dtype: int64
  - name: min_issues_commit_sha
    dtype: string
  - name: min_forks_path
    dtype: string
  - name: max_issues_owner
    dtype: string
  - name: max_issues_repo_id
    dtype: float64
  - name: stars
    dtype: int64
  - name: old_content
    dtype: string
  - name: min_stars_main_language
    dtype: string
  - name: min_stars_name
    dtype: string
  - name: max_stars
    dtype: int64
  - name: commit_sha
    dtype: string
  - name: min_issues
    dtype: float64
  - name: min_issues_languages_distribution
    dtype: string
  - name: size
    dtype: int64
  - name: comment_lines
    dtype: int64
  - name: max_issues_path
    dtype: string
  - name: max_stars_repo_id
    dtype: int64
  - name: min_forks_owner
    dtype: string
  - name: max_forks_owner
    dtype: string
  - name: min_issues_path
    dtype: string
  - name: min_len
    dtype: 'null'
  - name: max_forks_languages_distribution
    dtype: string
  - name: max_issues
    dtype: float64
  - name: min_forks_main_language
    dtype: string
  - name: median_len
    dtype: float64
  - name: min_stars_path
    dtype: string
  - name: min_forks_languages_distribution
    dtype: string
  - name: name
    dtype: string
  splits:
  - name: train
    num_bytes: 174601557.889927
    num_examples: 25000
  download_size: 69469316
  dataset_size: 174601557.889927
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---


# Dataset Summary
The dataset contains 25000 Kotlin code samples selected from [KStack](https://huggingface.co/datasets/JetBrains/KStack) dataset. The selection is performed based on the value of the code for learning algorithmic concepts in Kotlin. In total, the dataset contains about 23M [CodeLlama-7b](https://huggingface.co/codellama/CodeLlama-7b-hf) tokens (vocab size 32016).

# Dataset Collection Procedure
The filtering is performed using zero-shot quality estimation based on [Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2). The model is prompted to determine which of two files has higher "educational value for learning algorithms in Kotlin". The results of the comparisons are averaged and used to train a binary classifier based on [CodeT5p-220m](https://huggingface.co/Salesforce/codet5p-220m). The binary classifier is then applied to the entire KStack to obtain scores for each sample in the dataset. The log-probability of the classifier prediction used as a criterion of the selection.

# Opt-out
If you want your data to be removed from dataset, or have any other questions, please reach out to Sergey Titov [email protected]