Rosetta-Ko Suite
Collection
33 items • Updated
You need to agree to share your contact information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
By clicking "Agree" I confirm I have read and agree to the PoSTMEDIA Data Agreement for Model Training and agree that I intend to use this data for model training purposes only.
Log in or Sign Up to review the conditions and access this dataset content.
rosetta-ko-code-synth-dpo
Korean-native coding data — problems with solutions whose unit tests were actually executed and passed (execution-grounded rejection sampling).
Synthetic data generated with the Qwen3.6-27B teacher model — part of the Rosetta-KO suite for the Rosetta Korean LLM (PoSTMEDIA).
Code Suite
Sibling datasets from the same pipeline (each a separate repo):
| repo | format |
|---|---|
rosetta-ko-code-synth-sft |
supervised fine-tuning (user/assistant messages) |
rosetta-ko-code-synth-sft-think |
reasoning SFT (<think> chain-of-thought) |
rosetta-ko-code-synth-dpo (this repo) |
preference pairs (chosen/rejected) |
rosetta-ko-code-synth-dpo-think |
reasoning preference pairs |
rosetta-ko-code-synth-rlvr |
reinforcement-learning-with-verifiable-reward rows |
Data Fields
{"id": "...", "source_dataset": "...", "domain": "code",
"prompt": "...",
"chosen": [{"role": "assistant", "content": "..."}],
"rejected": [{"role": "assistant", "content": "..."}]}
prompt(string) — the Korean questionchosen/rejected— assistant responses;chosenis the verified-correct / better-grounded answer
Usage
from datasets import load_dataset
ds = load_dataset("PoSTMEDIA/rosetta-ko-code-synth-dpo", split="train")
License
Released under the Apache License 2.0.
Seed sources:
Citation
@misc{rosetta_ko_code_synth,
title = {Rosetta-KO Code Synthetic Dataset},
author = {PoSTMEDIA},
year = {2026},
howpublished = {Hugging Face: PoSTMEDIA/rosetta-ko-code-synth-dpo}
}
Contact
PoSTMEDIA — https://huggingface.co/PoSTMEDIA
- Downloads last month
- 19