datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Vietnam_HighSchool_Exam_Dataset
VNHSGE: VietNamese High School Graduation Examination Dataset for Large Language Models
Dataset released as part of the paper : https://arxiv.org/abs/2305.12199
Dataset source : https://github.com/Xdao85/VNHSGE/tree/main/Dataset/VNHSGE-V
Sample Size : 6663
MCQ questions with answers, covering 8 subjects
Literature subject was excluded as it wasnt in MCQ format and had essay like questions
high-school-physicstw-highschool-chat
Dataset Card for tw-highschool-chat
本資料集是以臺灣高中學科教材／知識為基礎合成的繁中對話集，每筆樣本含 <think> 思考過程與最終解答，可作為訓練具備臺灣高中知識回答能力之 chatbot／解題助理的 SFT 資料。
Dataset Details
Dataset Description
本資料集為 tw-highschool 的對話版本。每筆樣本以教材段落為 seed，由 LLM 生成「先思考、再回答」的單輪 QA。題型涵蓋觀念解釋、公式推導、選擇題、計算題等。
資料夾依學科分檔（例如 data/chemist.jsonl、data/math.jsonl 等），讓使用者可依領域取樣。
Curated by: Huang Liang Hsun
Language(s) (NLP): Traditional Chinese
License: cc-by-nc-sa-4.0
Dataset Sources
Repository:… See the full description on the dataset page: https://huggingface.co/datasets/lianghsun/tw-highschool-chat.qwen3_8b_chimera_fixedtopics_highschool_75k_probval_instill_n8_valredundancy5_round1qwen3_8b_chimera_fixedtopics_highschool_75k_nofilter_instill_n8_valredundancy5_round1highschool-math-reasoning-rl
High School Math Reasoning RL
Synthetic high school math reasoning data with train/test splits for SFT-style chat training.
Splits
train: 400 examples
test: 100 examples
Format
Each row contains:
messages: chat-style user and assistant turns
answer: a separate short-answer column
The assistant message contains the reasoning inside <think> ... </think> tags followed by the worked full_answer, which ends with Final answer: \boxed{...}.
highschool-math-reasoning
High School Math Reasoning
Synthetic high school math SFT data with train/test splits.
Format
Each row contains:
messages: chat-style user/assistant turns for SFT
answer: a separate short-answer column
The assistant message contains the reasoning inside <think> ... </think> tags followed by the worked full_answer, which ends with Final answer: \boxed{...}.
qwen3_8b_chimera_fixedtopics_highschool_75k_questions_nofilterqwen3_8b_chimera_fixedtopics_highschool_75k_questions_probvalmmlu-highschooltw-highschool
Dataset Card for tw-highschool
本資料集收錄臺灣高中四大主要學科（生物、化學、數學、物理）之繁體中文知識文本，作為訓練具備臺灣高中學科知識能力之繁中模型的補充語料。
Dataset Details
Dataset Description
資料以「學科」為 config 拆分：
biology：生物
chemistry：化學
math：數學
physics：物理
每個 config 內含對應學科之知識性段落（涵蓋觀念解說、公式推導、範例題、術語對照等），以繁體中文書面語為主。可作為高中學科 chatbot、教學助理之預訓練／微調補強。
Curated by: Huang Liang Hsun
Language(s) (NLP): Traditional Chinese
License: cc-by-nc-sa-4.0
Dataset Sources
Repository: lianghsun/tw-highschool
Chat… See the full description on the dataset page: https://huggingface.co/datasets/lianghsun/tw-highschool.HighschoolBiologyCleanedHighschool