Japanese translation of the GSM‑8K dataset, refined by post‑processing to restore missing “#### [回答]” markers in train/test splits and produced as part of a Tinker API LLM reinforcement learning experiment.
Description in other languages
Statistics
Statistics as of Nov 16, 2026
Hugging Face downloads (last 30 days)
108
Hugging Face likes
2
Listed
Sep 14, 2026
Popular in the same category
- ClimbLab-JaKantaHayashiAI15kClimbLab‑Ja is a 300 billion‑token Japanese corpus, filtered from LLM‑jp Corpus v4 into 20 clusters for pre‑training language models, released under ODC‑BY with original source licenses and stored in parquet format.
- llm-course-final-2026bumbuboon13kJapanese Wikipedia‑based RAG dataset for the LLM course final competition, containing ~1 million text chunks and a search index.
- JMedBenchColdog23331kJMedBench is a Japanese biomedical LLM benchmark comprising 20 datasets across five tasks (MCQA, NER, STS, etc.) sourced from MedMCQA, PubMedQA, MMLU, and others, each with its own license, and includes a note that translations may contain biases requiring human review.
- vntl-leaderboardlmg-anon857A leaderboard that ranks large language models by their semantic accuracy in translating Japanese visual novels into English, using cosine similarity over 256 samples and reporting chrF scores for comparison with tools like Sugoi Translator, Google Translate, Naver Papago, and Alibaba Translate.
- auto-wiki-qacl-nagoya747AutoWikiQA is the largest free Japanese QA dataset, providing over 2.3 million handcrafted‑filtered question–answer pairs automatically generated from Wikipedia with Swallow‑MX and LLMs (without rule‑based templates) to support knowledge‑teaching and retrieval‑augmented generation applications.
- jhumanevalkogi-jwu731JHumanEval is a hand‑translated Japanese version of the HumanEval benchmark, providing 164 Python programming problems with parallel English and Japanese comments to evaluate Japanese‑LLM code generation while preserving the original English errors.
Resources in similar research fields
- AnswerCarefullyllm-jp644AnswerCarefully Dataset supplies Japanese and multilingual data for commercial or non‑commercial LLM safety enhancement, prohibits any other use—including safety circumvention—allows derivative works with attribution, and carries a creator disclaimer of non‑liability for harms or service changes.Shared: Large language models・Corpora & datasets
- llm-japanese-datasetizumi-lab572Japanese instruction‑chat dataset for fine‑tuning LLMs (e.g., with LoRA), 9 M+ samples, recently updated to drop licensed Alpaca data, clean Wikipedia and ALT outputs, and released under CC‑BY‑SA 4.0.Shared: Large language models・Corpora & datasets
- oasst2-33k-jallm-jp551LLM‑jp provides a Japanese instruction‑tuning dataset, DeepL‑translated from an English subset of oasst2 (derived from kunishou/oasst2‑135k‑ja) and compiled by Kiyomaru and Kodama.Shared: Large language models・Corpora & datasets
- JGLUEllm-book546Dataset card for the JGLUE dataset used in the book “Large Language Model Introduction,” sourced from the original repo, with code licensed CC BY‑SA 4.0, data under the distributor’s license, citing Kurihara & Kawahara (in Japanese), and built on Shunsuke Kitada’s repository.Shared: Large language models・Corpora & datasets
- scaling-data-constrained-llmsllm-jp402A collection of Japanese and English web corpora—including a 9 B‑token Japanese set (JA‑WEB‑9B), 63 B‑token English and Japanese sets (EN‑WEB‑63B, JA‑WEB‑63B) and synthetic versions such as paraphrased JA‑PARAPHRASE‑63B, instruction‑style JA‑INSTRUCT‑63B, and translated JA‑TRANSLATE‑63B—used for studying data augmentation in pre‑training Japanese LLMs under data‑constrained settings.Shared: Large language models・Corpora & datasets
- kaken-trans-ja-enhpprc358Japanese text from the kaken subset of llm-jp-corpus-v3 translated into English with Qwen/Qwen2.5‑32B‑Instruct, released as an open Japanese–English parallel corpus under a CC‑BY 4.0 license that inherits the original dataset’s terms.Shared: Large language models・Corpora & datasets