datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Knowledge-QA-SingleTurn-Dataset
Knowledge QA Single-turn Dataset(知識質問データセット・シングルターン)
概要
本データセットは、Aratako/Synthetic-JP-Conversations-Magpie-Nemotron-4-10k から質問を抽出し、DeepSeek V3.2で整形、Kimi K2.5で回答を生成した シングルターンの知識質問応答データセット です。Reasoning有効化により思考過程も最終データに含まれ、質問の難易度に応じてReasoning effortが動的に切り替わります。
生成にはSDG-LOOMという合成データ生成パイプラインを用いました。(sdg-loom)
データの説明
項目
内容
件数
約7,000件
形式
JSONL(1行1JSON)
言語
日本語
ターン数
1ターン(質問1 + 回答1)
ソースデータセット… See the full description on the dataset page: https://huggingface.co/datasets/DataPilot/Knowledge-QA-SingleTurn-Dataset.hh-rlhf-49k-ja-single-turnThis dataset was created by automatically translating part of "Anthropic/hh-rlhf" into Japanese, and selected for single turn conversations.You can use this dataset for RLHF and DPO.
hh-rlhf repository
https://github.com/anthropics/hh-rlhf
Anthropic/hh-rlhf
https://huggingface.co/datasets/Anthropic/hh-rlhf
Chinese-Roleplay-SingleTurn请注意,个人模型经过characterEval的reward model进行DPO训练,因此使用本数据集进行SFT的模型在该榜单上会存在bias,导致分数异常偏高,请勿直接使用该榜单进行测试
简介
因已找到更优数据合成方案,为填充中文角色扮演数据集的空白,现开源部分中文角色扮演单轮对话数据集。
使用Refined-Anime-Text作为system prompt,使用小黄鸡随机query作为输入,调用个人角色扮演模型作为输出。
已处理为alpaca数据格式,方便大家处理和训练。经过验证,仅使用该数据集进行Lora微调即可获取一个效果还不错的模型~
chatGPT对比
character
question
answer_us
answer_chatGPT
黑须彼方是(省略……)黑须彼方有着许多有趣的爱好和特点。她是一个有点毒舌的人,但总能犀利地指出问题所在。她有着敏锐的洞察力,擅长看透人心。她经常以此来捉弄加贺正午。她与正午有着相同的口癖,张扬的性格(省略……)她的个性和爱好使她成为一个备受喜爱的角色。… See the full description on the dataset page: https://huggingface.co/datasets/LooksJuicy/Chinese-Roleplay-SingleTurn.hermis_singleTurnNepali_educationalalpaca_translate_en_then_answer_single_turn
Overview
This dataset is a manually constructed sequential instruction (translate-then-answer) tuning dataset derived from the famous Alpaca dataset.
Dataset features/keys
conversations - The user and assistant dialog turns formatted in a list.
dataset - Name of the dataset.
lang - Language(s) of the content in the format of l1-l2. In detail: l1 is the language of the instruction; l2 (always en) is the language of the translation and output.
task - chat.
split - train… See the full description on the dataset page: https://huggingface.co/datasets/pinzhenchen/alpaca_translate_en_then_answer_single_turn.open_perfectblend_singleturnnepali-json-mode-singleturn
This dataset is a remastered version of this dataset prepared using Adaption's Adaptive Data platform.
adaption-json_schema_instruction_data
This dataset contains instruction-following conversations where models are tasked with generating valid JSON objects based on provided schemas and user prompts. The samples cover diverse domains including healthcare, project management, engineering, and environmental science, requiring strict adherence to defined data structures. Each… See the full description on the dataset page: https://huggingface.co/datasets/himalaya-ai/nepali-json-mode-singleturn.singleturn-prd-conversationDans-Codemaxx-CodeFeedback-SingleTurn40k_mixed_singleturn_2femdom_singleturncodefeedback-single-turn-reformat10k_mixed_singleturn_2_1Hydrus-Next-Coder-Single-turn20k_mixed_singleturn_1single_turn_chatclaude_3.5s_single_turn_unslop_filtered-KTOSloPreferenceShareGPT40k_mixed_singleturn_1Mix of 40,000 short, single-turn prompts combined from OpenRLHF/prompt-collection-v0.1 and Babelscape/ALERT
20k_mixed_singleturn_210k_mixed_singleturn_2_2dpo_single_turn_20250903dolci-wildchat-think-singleturndolci-wildchat-think-singleturn-filteredsft-v2-merged-singleturn-datadadtalk-stt-single-turndadtalk-stt-single-turn_v2sft-v1-singleturn-ads-creativity
