Dataset · code-evol-instruct

Code Evol-Instruct (WizardCoder)

Builder: WizardLM team (WizardCoder authors)
Availability: unknown · checked 2026-09-24 · source
The WizardLMTeam HF org lists only WizardLM_evol_instruct_70k and _V2_196k, neither described as the code set. Third-party reproductions exist on the Hub (e.g. nickrosh/Evol-Instruct-Code-80k-v1); they are not the builders' data and are not recorded here. Whether the builders released their set elsewhere (e.g. GitHub) was not checked.

Raw record: /data/datasets/code-evol-instruct.json

Fields

id

code-evol-instruct

identifiers

huggingface

builder

WizardLM team (WizardCoder authors)

release_date

2023-06-14partial · source

note:  arXiv v1 date of the WizardCoder paper; no separate dataset release date.

availability

unknown · checked 2026-09-24 · source

note:  The WizardLMTeam HF org lists only WizardLM_evol_instruct_70k and _V2_196k, neither described as the code set. Third-party reproductions exist on the Hub (e.g. nickrosh/Evol-Instruct-Code-80k-v1); they are not the builders' data and are not recorded here. Whether the builders released their set elsewhere (e.g. GitHub) was not checked.

content

Paper: Code Alpaca ('around 20k samples') evolved iteratively with Code Evol-Instruct; 'OpenAI's gpt3.5turbo is used to evolve the dataset and generate responses. The evolved dataset consists of approximately 78k samples.' Used to fine-tune StarCoder and CodeLlama-34B-Python. The seed (Code Alpaca) is a dataset-to-dataset derivation, recorded here rather than as an edge.recorded · source

primary_sources

https://arxiv.org/abs/2306.08568

record_history

date:2026-09-24 · change:created from primary sources (Evol-Instruct chase) · by:wilson-pruitt + claude ·

Parents

Influence without weights

• distilled_from_outputs → ChatGPT (Nov. 2022 launch model) declared source Builders' own paper: 'OpenAI's gpt3.5turbo is used to evolve the dataset and generate responses.' Snapshot not_recorded.

Children

Training data

• ← trained_on WizardCoder-15B-V1.0 declared source Card: 'trained with 78k evolved code instructions'; paper: Code Alpaca evolved with Code Evol-Instruct, then StarCoder fine-tuned. The `code-evol-instruct` record has availability `unknown` (only third-party reproductions found); this edge is to that record, not to a released file.

Read in

No station on the reading path has touched this record yet.