instruction stringlengths 53 190 | input stringlengths 288 1.05k | output stringlengths 350 862 | source stringclasses 1
value | original_id stringlengths 8 44 |
|---|---|---|---|---|
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: prompts.chat
Description: a.k.a. Awesome ChatGPT Prompts. This is a Dataset Repository mirror of prompts.chat — a social platform for AI prompts. prompts.chat is an open-source platform where users can share, discover, and collect AI prompts from the community.
Tags: task_categories:question-answering, task_categ... | prompts.chat is a Hugging Face mirror of the Awesome ChatGPT Prompts community repository. It is a text dataset of shared AI prompts intended for prompt engineering, chatbot design, question answering, and text generation workflows. The dataset is relatively small, in the 1K to 10K size range, and is distributed in CSV... | huggingface | fka/prompts.chat |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: fineweb
Description: FineWeb consists of more than 18.5T tokens of cleaned and deduplicated English web data from CommonCrawl. The data processing pipeline is optimized for LLM performance and ran on the datatrove library. FineWeb was originally meant to be a fully open replication of RefinedWeb.
Tags: task_categ... | FineWeb is a massive English web corpus designed for large-scale language model pretraining. It contains over 18.5 trillion tokens of cleaned and deduplicated CommonCrawl text and is optimized for LLM performance. The dataset is extremely large, tagged in the 10B<n<100B size category, and supports text generation use c... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: hh-rlhf
Description: This repository provides access to human preference data about helpfulness and harmlessness from Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. These data are meant to train preference or reward models for subsequent RLHF training and are not meant ... | hh-rlhf is a human feedback dataset created for reinforcement learning from human feedback pipelines, especially for modeling helpfulness and harmlessness preferences. It is best suited for reward model training, preference learning, and alignment research rather than direct supervised fine-tuning of chat assistants. T... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: OpenOrca
Description: The OpenOrca dataset is a rich collection of augmented FLAN data aligned as closely as possible with the distributions outlined in the Orca paper. It has been instrumental in generating high-performing model checkpoints and is intended as a valuable resource for NLP researchers and developer... | OpenOrca is a large English instruction-style dataset built from augmented FLAN data and designed to reflect the distributions discussed in the Orca work. It supports a broad range of NLP and LLM tasks including text generation, question answering, summarization, classification, and feature extraction. The dataset is t... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: oasst1
Description: OpenAssistant Conversations Dataset (OASST1) is a human-generated, human-annotated assistant-style conversation corpus consisting of 161,443 messages in 35 different languages, annotated with 461,292 quality ratings and over 10,000 fully annotated conversation trees. It is a product of a world... | OASST1 is a multilingual assistant conversation dataset created through large-scale human generation and annotation. It contains over 161K messages across 35 languages, along with extensive quality ratings and annotated conversation trees, making it highly valuable for supervised fine-tuning, dialogue modeling, multili... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: gsm8k
Description: Dataset Card for GSM8K. GSM8K (Grade School Math 8K) is a dataset of 8.5K high quality linguistically diverse grade school math word problems. The dataset was created to support the task of question answering on basic mathematical problems that require multi-step reasoning. These problems take ... | GSM8K is a popular English text dataset of grade-school math word problems designed for multi-step reasoning and question answering. It contains original, crowdsourced problems that typically require 2 to 8 arithmetic steps to solve. The dataset is widely used as an official benchmark for math reasoning, is distributed... | huggingface | openai/gsm8k |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: wikipedia
Description: Dataset Card for Wikimedia Wikipedia. Wikipedia dataset containing cleaned articles of all languages. The dataset is built from the Wikipedia dumps with one subset per language, each containing a single train split. Each example contains the content of one full Wikipedia article with cleani... | Wikimedia Wikipedia is a massive multilingual text dataset containing cleaned full articles from Wikipedia dumps, organized by language with one train split per language. It supports text generation, language modeling, and masked language modeling, and spans hundreds of languages. The dataset is distributed in parquet ... | huggingface | wikimedia/wikipedia |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: EasyNegative
Description: Negative Embedding. This is a Negative Embedding trained with Counterfeit. Please use it in the stable-diffusion-webui embeddings folder. It can be used with other models, but the effectiveness is not certain.
Tags: license:other, size_categories:n<1K, format:imagefolder, modality:image,... | EasyNegative is a small image-related resource described as a negative embedding trained with Counterfeit for use in Stable Diffusion WebUI embeddings. It is intended for image generation workflows rather than text tasks, and may also be used with other models though effectiveness is uncertain. The resource is very sma... | huggingface | gsdf/EasyNegative |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: RedPajama-Data-1T
Description: RedPajama is a clean-room, fully open-source implementation of the LLaMa dataset.
Tags: task_categories:text-generation, language:en, size_categories:1M<n<10M, modality:text, library:datasets, library:mlcroissant, region:us
Downloads: 2123
Likes: 1151 | RedPajama-Data-1T is an English text dataset for text generation, presented as a clean-room, fully open-source implementation of the dataset used to train LLaMA-style models. It is aimed at large-scale language model pretraining and open research. Despite relatively modest download counts of 2,123, it has strong commun... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: medical-o1-reasoning-SFT
Description: News. [2025/04/22] We split the data and kept only the medical SFT dataset. [2025/02/22] We released the distilled dataset from Deepseek-R1 based on medical verifiable problems. [2024/12/25] We open-sourced the medical reasoning dataset for SFT, built on medical verifiable pr... | medical-o1-reasoning-SFT is a bilingual English-Chinese text dataset for supervised fine-tuning on medical reasoning and question answering. It is built around medical verifiable problems and includes distilled reasoning data derived from DeepSeek-R1-style outputs. The dataset is released in JSON format under Apache-2.... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: fineweb-edu
Description: 1.3 trillion tokens of educational web data filtered from FineWeb using an educational quality classifier. This is the 1.3T token version.
Tags: task_categories:text-generation, language:en, license:odc-by, size_categories:1B<n<10B, format:parquet, modality:tabular, modality:text, library... | FineWeb-Edu is a large-scale English educational web dataset for text generation and LLM pretraining. It contains about 1.3 trillion tokens filtered from FineWeb using an educational quality classifier, making it especially suitable for training models on higher-quality instructional and knowledge-rich web text. The da... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: dolma
Description: Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research
Tags: task_categories:text-generation, language:en, license:odc-by, size_categories:n>1T, arxiv:2402.00159, arxiv:2301.13688, region:us, language-modeling, casual-lm, llm
Downloads: 3087
Likes: 1021 | Dolma is an open English corpus designed specifically for language model pretraining research, with a reported scale of three trillion tokens. It is oriented toward text generation and language modeling use cases and is distributed under the ODC-BY license. The dataset is notable for its explicit positioning as a resea... | huggingface | allenai/dolma |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: the-stack
Description: The Stack is a near-deduplicated code dataset. Initial release included 30 programming languages and permissive licenses; later versions excluded weak copyleft licenses and expanded permissive license coverage. The resulting dataset is around 3TB in size.
Tags: task_categories:text-generati... | The Stack is a large multilingual code dataset built for text generation and code model pretraining. It consists of near-deduplicated source code spanning many programming languages and emphasizes permissive licensing, with later releases removing weak copyleft licenses and expanding acceptable license coverage. The da... | huggingface | bigcode/the-stack |
Summarize the dataset for LLM pretraining or fine-tuning use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: TinyStories
Description: Dataset containing synthetically generated short stories by GPT-3.5 and GPT-4 using a small vocabulary. Used in the TinyStories paper, with train and validation files provided.
Tags: task_categories:text-generation, language:en, license:cdla-sharing-1.0, size_categories:1M<n<10M, format:p... | TinyStories is a small English text generation dataset made of synthetic short stories generated by GPT-3.5 and GPT-4 with intentionally simple vocabulary. Released under the CDLA-Sharing 1.0 license, it is well suited for lightweight language modeling experiments, curriculum learning, and research on small models or l... | huggingface | roneneldan/TinyStories |
Summarize the dataset for instruction tuning use, including its scale, tasks, language, license, notable metadata, and why it may be useful. | Name: databricks-dolly-15k
Description: Open source dataset of instruction-following records generated by Databricks employees across categories such as brainstorming, classification, closed QA, generation, information extraction, open QA, and summarization.
Tags: task_categories:question-answering, task_categories:sum... | databricks-dolly-15k is an English instruction-following dataset intended for supervised fine-tuning of assistant models. It contains about 15,000 human-generated examples spanning tasks such as brainstorming, classification, question answering, generation, information extraction, and summarization. The dataset is dist... | huggingface | databricks/databricks-dolly-15k |
Summarize the dataset and identify its primary use case, modality, language, license, approximate size, and popularity. | Name: alpaca
Description: Alpaca is a dataset of 52,000 instructions and demonstrations generated by OpenAI's text-davinci-003 engine. This instruction data can be used to conduct instruction-tuning for language models and make the language model follow instruction better.
Tags: task_categories:text-generation, languag... | Alpaca is an English text dataset for instruction tuning and text generation. It contains about 52,000 instruction-response demonstrations generated with text-davinci-003 and is intended to help language models better follow instructions. The dataset uses the CC-BY-NC-4.0 license, is distributed in Parquet format, and ... | huggingface | tatsu-lab/alpaca |
Summarize the dataset and identify its purpose, modality, language, license, approximate size, and popularity. | Name: bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to somehow train the negative prompt as an embedding, thus unifying the basis of the negative prompt into one word or embedding. Embedding has proven to be very helpful for the generation of hands. To use this embed... | bad_prompt is an English image-related dataset/resource associated with Stable Diffusion and textual inversion. Its purpose is to provide a negative embedding that can be used in generation workflows, including text-to-image and image-to-image tasks, to improve results such as hand generation. It is licensed under Crea... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset and list its key metadata including source, name, description, tags, downloads, likes, and license. | Dataset record: source=huggingface; id=tiiuae/falcon-refinedweb; name=falcon-refinedweb; description=Falcon RefinedWeb is a massive English web dataset built by TII and released under an ODC-By 1.0 license. RefinedWeb is built through stringent filtering and large-scale deduplication of CommonCrawl; models trained on R... | Falcon RefinedWeb is a large English web dataset from TII hosted on Hugging Face. It is designed for text generation use cases and was created from CommonCrawl using strong filtering and large-scale deduplication. The dataset is notable for achieving strong training performance using only web data and for being multimo... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset and list its key metadata including source, name, description, tags, downloads, likes, and notable characteristics. | Dataset record: source=huggingface; id=lmsys/lmsys-chat-1m; name=lmsys-chat-1m; description=LMSYS-Chat-1M is a large-scale real-world LLM conversation dataset containing one million conversations with 25 state-of-the-art LLMs. It was collected from 210K unique IP addresses on the Vicuna demo and Chatbot Arena website f... | LMSYS-Chat-1M is a real-world conversation dataset hosted on Hugging Face that contains one million chats involving 25 modern LLMs. The data was gathered from Vicuna demo and Chatbot Arena traffic between April and August 2023, covering 210K unique IP addresses. Each record includes conversation ID, model name, convers... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise description for model training relevance. | Name: ShareGPT_Vicuna_unfiltered
Description: Further cleaning done. Please look through the dataset and ensure that I didn't miss anything. Update: Confirmed working method for training the model. Two choices: removes instances of "I'm sorry, but" or keeps them.
Tags: language:en, license:apache-2.0, region:us
Downloa... | ShareGPT_Vicuna_unfiltered is an English-language Hugging Face dataset intended for conversational model training and data cleaning workflows. It includes cleaned ShareGPT/Vicuna-style data, with alternative versions that either remove or retain common refusal phrasing such as "I'm sorry, but". The dataset is licensed ... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise description for model training relevance. | Name: finepdfs
Description: Liberating 3T of the finest tokens from PDFs. FinePDFs is the largest publicly available corpus sourced exclusively from PDFs, containing about 3 trillion tokens across 475 million documents in 1733 languages.
Tags: task_categories:text-generation, many language tags, license:odc-by, size_ca... | finepdfs is a massive multilingual text-generation dataset on Hugging Face built exclusively from PDF sources. It contains approximately 3 trillion tokens from 475 million documents spanning 1,733 languages, making it a significant corpus for large-scale pretraining, multilingual modeling, and document-derived text res... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset for a machine learning practitioner, including its purpose, modality, scale, collection or labeling details, and notable usage constraints or signals from the metadata. | Name: PhysicalAI-Autonomous-Vehicles
Description: PHYSICAL AI AUTONOMOUS VEHICLES
The PhysicalAI-Autonomous-Vehicles dataset provides one of the largest, geographically diverse collections of multi-sensor data empowering AV researchers to build the next generation of Physical AI based end-to-end driving systems. This ... | PhysicalAI-Autonomous-Vehicles is a large-scale autonomous driving dataset focused on Physical AI and end-to-end driving research. It contains geographically diverse multi-sensor driving data and is intended to support development of next-generation AV systems. The dataset includes about 1700 hours of driving, with bot... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset for a machine learning practitioner, including its purpose, modality, scale, subset information, and notable metadata such as license, format, and domain coverage. | Name: OpenThoughts-114k
Description: Open-Thoughts-114k
Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles.
Available Subsets: default subset containing ready-to-train data used to finetune the OpenThinker-7B and OpenThinker-32B models.
Tags: license:apache-2.0, ... | OpenThoughts-114k is a synthetic text reasoning dataset designed for training and fine-tuning language models on complex reasoning tasks. It contains 114,000 high-quality examples spanning math, science, coding, and puzzles. The default subset is ready-to-train and was used to finetune the OpenThinker-7B and OpenThinke... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction tuning, including its purpose, scale, modality, notable characteristics, and popularity signals. | Name: OpenHermes-2.5
Source: huggingface
Dataset ID: teknium/OpenHermes-2.5
Description: This is the dataset that made OpenHermes 2.5 and Nous Hermes 2 series of models. It is described as the exact compilation and curation of many open source datasets and custom created synthetic datasets used to underpin recent advan... | OpenHermes-2.5 is a large-scale English text dataset on Hugging Face designed for instruction tuning and LLM development. It is a JSON-formatted compilation of many open-source datasets along with custom synthetic data, and is presented as the training foundation for the OpenHermes 2.5 and Nous Hermes 2 model series. T... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction tuning, including its purpose, data quality focus, scale, format, and popularity signals. | Name: alpaca-cleaned
Source: huggingface
Dataset ID: yahma/alpaca-cleaned
Description: This is a cleaned version of the original Alpaca Dataset released by Stanford. It fixes issues identified in the original release, including hallucinations caused by instructions that referenced internet data.
Tags: task_categories:t... | alpaca-cleaned is an English text-generation dataset on Hugging Face intended for instruction fine-tuning. It is a cleaned version of Stanford's original Alpaca dataset, created to improve data quality by fixing known issues such as hallucination-prone instructions that referenced unavailable internet information. The ... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset based on the provided metadata. | Source: huggingface
ID: cais/hle
Name: hle
Description: Humanity's Last Exam (HLE) is a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. It consists of 2,500 questions across dozens of subjects. Important note: do... | hle is an official Hugging Face benchmark dataset from cais/hle. It is a multimodal academic benchmark containing text and image data, distributed in Parquet format under the MIT license. The dataset is positioned as "Humanity's Last Exam," with 2,500 closed-ended questions spanning many subjects at the frontier of hum... | huggingface | cais/hle |
Summarize the dataset based on the provided metadata. | Source: huggingface
ID: HuggingFaceFW/fineweb-2
Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining data for over 1000 languages. It is fully reproducible, available under the permissive ODC-By 1.0 license, and validated through hundreds of ablation e... | fineweb-2 is a large-scale Hugging Face dataset from HuggingFaceFW designed for text generation and multilingual pretraining. It is the second iteration of FineWeb and provides high-quality, fully reproducible pretraining data across more than 1000 languages. The dataset is licensed under ODC-By, includes text and tabu... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset based on its metadata, including its purpose, modality, likely use cases, language, license, scale, and notable popularity signals. | Name: imagenet-1k
Source: huggingface
Dataset ID: ILSVRC/imagenet-1k
Description: ILSVRC 2012, commonly known as 'ImageNet' is an image dataset organized according to the WordNet hierarchy. Each meaningful concept in WordNet, possibly described by multiple words or word phrases, is called a "synonym set" or "synset". T... | ImageNet-1k is a large-scale image classification dataset hosted on Hugging Face, based on the ILSVRC 2012 benchmark. It contains images organized by WordNet synsets and is designed for multi-class image classification. The dataset modality is image, the primary language metadata is English, and it is monolingual. It a... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset based on its metadata, including its purpose, content type, languages, license, and popularity. | Name: Alpaca-CoT
Source: huggingface
Dataset ID: QingyiSi/Alpaca-CoT
Description: This repository will continuously collect various instruction tuning datasets. And we standardize different datasets into the same format, which can be directly loaded by the code of Alpaca model. We also have conducted empirical study on... | Alpaca-CoT is an instruction-finetuning dataset collection intended for training or studying Alpaca-style language models. It aggregates various instruction-tuning datasets and standardizes them into a unified format compatible with Alpaca model code. Based on the tags, it supports multiple languages, including English... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset and identify its main purpose, modalities, languages, tasks, license, and notable characteristics. | Name: PersonaHub
Description: Scaling Synthetic Data Creation with 1,000,000,000 Personas. This repo releases data introduced in the paper "Scaling Synthetic Data Creation with 1,000,000,000 Personas." It proposes a persona-driven data synthesis methodology that leverages various perspectives within a large language mo... | PersonaHub is a Hugging Face dataset focused on large-scale synthetic data creation using persona-driven generation. Its main purpose is to provide a massive collection of diverse personas that can be used to synthesize varied training data from different perspectives within an LLM. The dataset is text-based, provided ... | huggingface | proj-persona/PersonaHub |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: prompts.chat
Description: a.k.a. Awesome ChatGPT Prompts. This is a Dataset Repository mirror of prompts.chat — a social platform for AI prompts. prompts.chat is an open-source platform where users can share, discover, and collect AI prompts from the community.
Tags: task_categories:question-answering, task_categ... | prompts.chat is a Hugging Face mirror of the Awesome ChatGPT Prompts community repository. It is a text dataset of shared AI prompts intended for prompt engineering, chatbot design, question answering, and text generation workflows. The dataset is relatively small, in the 1K to 10K size range, and is distributed in CSV... | huggingface | fka/prompts.chat |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: fineweb
Description: FineWeb consists of more than 18.5T tokens of cleaned and deduplicated English web data from CommonCrawl. The data processing pipeline is optimized for LLM performance and ran on the datatrove library. FineWeb was originally meant to be a fully open replication of RefinedWeb.
Tags: task_categ... | FineWeb is a massive English web corpus designed for large-scale language model pretraining. It contains over 18.5 trillion tokens of cleaned and deduplicated CommonCrawl text and is optimized for LLM performance. The dataset is extremely large, tagged in the 10B<n<100B size category, and supports text generation use c... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: hh-rlhf
Description: This repository provides access to human preference data about helpfulness and harmlessness from Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. These data are meant to train preference or reward models for subsequent RLHF training and are not meant ... | hh-rlhf is a human feedback dataset created for reinforcement learning from human feedback pipelines, especially for modeling helpfulness and harmlessness preferences. It is best suited for reward model training, preference learning, and alignment research rather than direct supervised fine-tuning of chat assistants. T... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: OpenOrca
Description: The OpenOrca dataset is a rich collection of augmented FLAN data aligned as closely as possible with the distributions outlined in the Orca paper. It has been instrumental in generating high-performing model checkpoints and is intended as a valuable resource for NLP researchers and developer... | OpenOrca is a large English instruction-style dataset built from augmented FLAN data and designed to reflect the distributions discussed in the Orca work. It supports a broad range of NLP and LLM tasks including text generation, question answering, summarization, classification, and feature extraction. The dataset is t... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags. | Name: oasst1
Description: OpenAssistant Conversations Dataset (OASST1) is a human-generated, human-annotated assistant-style conversation corpus consisting of 161,443 messages in 35 different languages, annotated with 461,292 quality ratings and over 10,000 fully annotated conversation trees. It is a product of a world... | OASST1 is a multilingual assistant conversation dataset created through large-scale human generation and annotation. It contains over 161K messages across 35 languages, along with extensive quality ratings and annotated conversation trees, making it highly valuable for supervised fine-tuning, dialogue modeling, multili... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: gsm8k
Description: Dataset Card for GSM8K. GSM8K (Grade School Math 8K) is a dataset of 8.5K high quality linguistically diverse grade school math word problems. The dataset was created to support the task of question answering on basic mathematical problems that require multi-step reasoning. These problems take ... | GSM8K is a popular English text dataset of grade-school math word problems designed for multi-step reasoning and question answering. It contains original, crowdsourced problems that typically require 2 to 8 arithmetic steps to solve. The dataset is widely used as an official benchmark for math reasoning, is distributed... | huggingface | openai/gsm8k |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: wikipedia
Description: Dataset Card for Wikimedia Wikipedia. Wikipedia dataset containing cleaned articles of all languages. The dataset is built from the Wikipedia dumps with one subset per language, each containing a single train split. Each example contains the content of one full Wikipedia article with cleani... | Wikimedia Wikipedia is a massive multilingual text dataset containing cleaned full articles from Wikipedia dumps, organized by language with one train split per language. It supports text generation, language modeling, and masked language modeling, and spans hundreds of languages. The dataset is distributed in parquet ... | huggingface | wikimedia/wikipedia |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: EasyNegative
Description: Negative Embedding. This is a Negative Embedding trained with Counterfeit. Please use it in the stable-diffusion-webui embeddings folder. It can be used with other models, but the effectiveness is not certain.
Tags: license:other, size_categories:n<1K, format:imagefolder, modality:image,... | EasyNegative is a small image-related resource described as a negative embedding trained with Counterfeit for use in Stable Diffusion WebUI embeddings. It is intended for image generation workflows rather than text tasks, and may also be used with other models though effectiveness is uncertain. The resource is very sma... | huggingface | gsdf/EasyNegative |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: RedPajama-Data-1T
Description: RedPajama is a clean-room, fully open-source implementation of the LLaMa dataset.
Tags: task_categories:text-generation, language:en, size_categories:1M<n<10M, modality:text, library:datasets, library:mlcroissant, region:us
Downloads: 2123
Likes: 1151 | RedPajama-Data-1T is an English text dataset for text generation, presented as a clean-room, fully open-source implementation of the dataset used to train LLaMA-style models. It is aimed at large-scale language model pretraining and open research. Despite relatively modest download counts of 2,123, it has strong commun... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset or resource in a concise way for LLM training, including its purpose, modality, language scope, notable tags, and popularity signals. | Name: medical-o1-reasoning-SFT
Description: News. [2025/04/22] We split the data and kept only the medical SFT dataset. [2025/02/22] We released the distilled dataset from Deepseek-R1 based on medical verifiable problems. [2024/12/25] We open-sourced the medical reasoning dataset for SFT, built on medical verifiable pr... | medical-o1-reasoning-SFT is a bilingual English-Chinese text dataset for supervised fine-tuning on medical reasoning and question answering. It is built around medical verifiable problems and includes distilled reasoning data derived from DeepSeek-R1-style outputs. The dataset is released in JSON format under Apache-2.... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: fineweb-edu
Description: 1.3 trillion tokens of educational web data filtered from FineWeb using an educational quality classifier. This is the 1.3T token version.
Tags: task_categories:text-generation, language:en, license:odc-by, size_categories:1B<n<10B, format:parquet, modality:tabular, modality:text, library... | FineWeb-Edu is a large-scale English educational web dataset for text generation and LLM pretraining. It contains about 1.3 trillion tokens filtered from FineWeb using an educational quality classifier, making it especially suitable for training models on higher-quality instructional and knowledge-rich web text. The da... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: dolma
Description: Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research
Tags: task_categories:text-generation, language:en, license:odc-by, size_categories:n>1T, arxiv:2402.00159, arxiv:2301.13688, region:us, language-modeling, casual-lm, llm
Downloads: 3087
Likes: 1021 | Dolma is an open English corpus designed specifically for language model pretraining research, with a reported scale of three trillion tokens. It is oriented toward text generation and language modeling use cases and is distributed under the ODC-BY license. The dataset is notable for its explicit positioning as a resea... | huggingface | allenai/dolma |
Summarize the dataset for LLM pretraining use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: the-stack
Description: The Stack is a near-deduplicated code dataset. Initial release included 30 programming languages and permissive licenses; later versions excluded weak copyleft licenses and expanded permissive license coverage. The resulting dataset is around 3TB in size.
Tags: task_categories:text-generati... | The Stack is a large multilingual code dataset built for text generation and code model pretraining. It consists of near-deduplicated source code spanning many programming languages and emphasizes permissive licensing, with later releases removing weak copyleft licenses and expanding acceptable license coverage. The da... | huggingface | bigcode/the-stack |
Summarize the dataset for LLM pretraining or fine-tuning use, including its scale, domain, language, license, notable metadata, and why it may be useful. | Name: TinyStories
Description: Dataset containing synthetically generated short stories by GPT-3.5 and GPT-4 using a small vocabulary. Used in the TinyStories paper, with train and validation files provided.
Tags: task_categories:text-generation, language:en, license:cdla-sharing-1.0, size_categories:1M<n<10M, format:p... | TinyStories is a small English text generation dataset made of synthetic short stories generated by GPT-3.5 and GPT-4 with intentionally simple vocabulary. Released under the CDLA-Sharing 1.0 license, it is well suited for lightweight language modeling experiments, curriculum learning, and research on small models or l... | huggingface | roneneldan/TinyStories |
Summarize the dataset for instruction tuning use, including its scale, tasks, language, license, notable metadata, and why it may be useful. | Name: databricks-dolly-15k
Description: Open source dataset of instruction-following records generated by Databricks employees across categories such as brainstorming, classification, closed QA, generation, information extraction, open QA, and summarization.
Tags: task_categories:question-answering, task_categories:sum... | databricks-dolly-15k is an English instruction-following dataset intended for supervised fine-tuning of assistant models. It contains about 15,000 human-generated examples spanning tasks such as brainstorming, classification, question answering, generation, information extraction, and summarization. The dataset is dist... | huggingface | databricks/databricks-dolly-15k |
Summarize the dataset and identify its primary use case, modality, language, license, approximate size, and popularity. | Name: alpaca
Description: Alpaca is a dataset of 52,000 instructions and demonstrations generated by OpenAI's text-davinci-003 engine. This instruction data can be used to conduct instruction-tuning for language models and make the language model follow instruction better.
Tags: task_categories:text-generation, languag... | Alpaca is an English text dataset for instruction tuning and text generation. It contains about 52,000 instruction-response demonstrations generated with text-davinci-003 and is intended to help language models better follow instructions. The dataset uses the CC-BY-NC-4.0 license, is distributed in Parquet format, and ... | huggingface | tatsu-lab/alpaca |
Summarize the dataset and identify its purpose, modality, language, license, approximate size, and popularity. | Name: bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to somehow train the negative prompt as an embedding, thus unifying the basis of the negative prompt into one word or embedding. Embedding has proven to be very helpful for the generation of hands. To use this embed... | bad_prompt is an English image-related dataset/resource associated with Stable Diffusion and textual inversion. Its purpose is to provide a negative embedding that can be used in generation workflows, including text-to-image and image-to-image tasks, to improve results such as hand generation. It is licensed under Crea... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset and list its key metadata including source, name, description, tags, downloads, likes, and license. | Dataset record: source=huggingface; id=tiiuae/falcon-refinedweb; name=falcon-refinedweb; description=Falcon RefinedWeb is a massive English web dataset built by TII and released under an ODC-By 1.0 license. RefinedWeb is built through stringent filtering and large-scale deduplication of CommonCrawl; models trained on R... | Falcon RefinedWeb is a large English web dataset from TII hosted on Hugging Face. It is designed for text generation use cases and was created from CommonCrawl using strong filtering and large-scale deduplication. The dataset is notable for achieving strong training performance using only web data and for being multimo... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset and list its key metadata including source, name, description, tags, downloads, likes, and notable characteristics. | Dataset record: source=huggingface; id=lmsys/lmsys-chat-1m; name=lmsys-chat-1m; description=LMSYS-Chat-1M is a large-scale real-world LLM conversation dataset containing one million conversations with 25 state-of-the-art LLMs. It was collected from 210K unique IP addresses on the Vicuna demo and Chatbot Arena website f... | LMSYS-Chat-1M is a real-world conversation dataset hosted on Hugging Face that contains one million chats involving 25 modern LLMs. The data was gathered from Vicuna demo and Chatbot Arena traffic between April and August 2023, covering 210K unique IP addresses. Each record includes conversation ID, model name, convers... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise description for model training relevance. | Name: ShareGPT_Vicuna_unfiltered
Description: Further cleaning done. Please look through the dataset and ensure that I didn't miss anything. Update: Confirmed working method for training the model. Two choices: removes instances of "I'm sorry, but" or keeps them.
Tags: language:en, license:apache-2.0, region:us
Downloa... | ShareGPT_Vicuna_unfiltered is an English-language Hugging Face dataset intended for conversational model training and data cleaning workflows. It includes cleaned ShareGPT/Vicuna-style data, with alternative versions that either remove or retain common refusal phrasing such as "I'm sorry, but". The dataset is licensed ... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise description for model training relevance. | Name: finepdfs
Description: Liberating 3T of the finest tokens from PDFs. FinePDFs is the largest publicly available corpus sourced exclusively from PDFs, containing about 3 trillion tokens across 475 million documents in 1733 languages.
Tags: task_categories:text-generation, many language tags, license:odc-by, size_ca... | finepdfs is a massive multilingual text-generation dataset on Hugging Face built exclusively from PDF sources. It contains approximately 3 trillion tokens from 475 million documents spanning 1,733 languages, making it a significant corpus for large-scale pretraining, multilingual modeling, and document-derived text res... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset for a machine learning practitioner, including its purpose, modality, scale, collection or labeling details, and notable usage constraints or signals from the metadata. | Name: PhysicalAI-Autonomous-Vehicles
Description: PHYSICAL AI AUTONOMOUS VEHICLES
The PhysicalAI-Autonomous-Vehicles dataset provides one of the largest, geographically diverse collections of multi-sensor data empowering AV researchers to build the next generation of Physical AI based end-to-end driving systems. This ... | PhysicalAI-Autonomous-Vehicles is a large-scale autonomous driving dataset focused on Physical AI and end-to-end driving research. It contains geographically diverse multi-sensor driving data and is intended to support development of next-generation AV systems. The dataset includes about 1700 hours of driving, with bot... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset for a machine learning practitioner, including its purpose, modality, scale, subset information, and notable metadata such as license, format, and domain coverage. | Name: OpenThoughts-114k
Description: Open-Thoughts-114k
Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles.
Available Subsets: default subset containing ready-to-train data used to finetune the OpenThinker-7B and OpenThinker-32B models.
Tags: license:apache-2.0, ... | OpenThoughts-114k is a synthetic text reasoning dataset designed for training and fine-tuning language models on complex reasoning tasks. It contains 114,000 high-quality examples spanning math, science, coding, and puzzles. The default subset is ready-to-train and was used to finetune the OpenThinker-7B and OpenThinke... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction tuning, including its purpose, scale, modality, notable characteristics, and popularity signals. | Name: OpenHermes-2.5
Source: huggingface
Dataset ID: teknium/OpenHermes-2.5
Description: This is the dataset that made OpenHermes 2.5 and Nous Hermes 2 series of models. It is described as the exact compilation and curation of many open source datasets and custom created synthetic datasets used to underpin recent advan... | OpenHermes-2.5 is a large-scale English text dataset on Hugging Face designed for instruction tuning and LLM development. It is a JSON-formatted compilation of many open-source datasets along with custom synthetic data, and is presented as the training foundation for the OpenHermes 2.5 and Nous Hermes 2 model series. T... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction tuning, including its purpose, data quality focus, scale, format, and popularity signals. | Name: alpaca-cleaned
Source: huggingface
Dataset ID: yahma/alpaca-cleaned
Description: This is a cleaned version of the original Alpaca Dataset released by Stanford. It fixes issues identified in the original release, including hallucinations caused by instructions that referenced internet data.
Tags: task_categories:t... | alpaca-cleaned is an English text-generation dataset on Hugging Face intended for instruction fine-tuning. It is a cleaned version of Stanford's original Alpaca dataset, created to improve data quality by fixing known issues such as hallucination-prone instructions that referenced unavailable internet information. The ... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset based on the provided metadata. | Source: huggingface
ID: cais/hle
Name: hle
Description: Humanity's Last Exam (HLE) is a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. It consists of 2,500 questions across dozens of subjects. Important note: do... | hle is an official Hugging Face benchmark dataset from cais/hle. It is a multimodal academic benchmark containing text and image data, distributed in Parquet format under the MIT license. The dataset is positioned as "Humanity's Last Exam," with 2,500 closed-ended questions spanning many subjects at the frontier of hum... | huggingface | cais/hle |
Summarize the dataset based on the provided metadata. | Source: huggingface
ID: HuggingFaceFW/fineweb-2
Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining data for over 1000 languages. It is fully reproducible, available under the permissive ODC-By 1.0 license, and validated through hundreds of ablation e... | fineweb-2 is a large-scale Hugging Face dataset from HuggingFaceFW designed for text generation and multilingual pretraining. It is the second iteration of FineWeb and provides high-quality, fully reproducible pretraining data across more than 1000 languages. The dataset is licensed under ODC-By, includes text and tabu... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset based on its metadata, including its purpose, modality, likely use cases, language, license, scale, and notable popularity signals. | Name: imagenet-1k
Source: huggingface
Dataset ID: ILSVRC/imagenet-1k
Description: ILSVRC 2012, commonly known as 'ImageNet' is an image dataset organized according to the WordNet hierarchy. Each meaningful concept in WordNet, possibly described by multiple words or word phrases, is called a "synonym set" or "synset". T... | ImageNet-1k is a large-scale image classification dataset hosted on Hugging Face, based on the ILSVRC 2012 benchmark. It contains images organized by WordNet synsets and is designed for multi-class image classification. The dataset modality is image, the primary language metadata is English, and it is monolingual. It a... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset based on its metadata, including its purpose, content type, languages, license, and popularity. | Name: Alpaca-CoT
Source: huggingface
Dataset ID: QingyiSi/Alpaca-CoT
Description: This repository will continuously collect various instruction tuning datasets. And we standardize different datasets into the same format, which can be directly loaded by the code of Alpaca model. We also have conducted empirical study on... | Alpaca-CoT is an instruction-finetuning dataset collection intended for training or studying Alpaca-style language models. It aggregates various instruction-tuning datasets and standardizes them into a unified format compatible with Alpaca model code. Based on the tags, it supports multiple languages, including English... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset and identify its main purpose, modalities, languages, tasks, license, and notable characteristics. | Name: PersonaHub
Description: Scaling Synthetic Data Creation with 1,000,000,000 Personas. This repo releases data introduced in the paper "Scaling Synthetic Data Creation with 1,000,000,000 Personas." It proposes a persona-driven data synthesis methodology that leverages various perspectives within a large language mo... | PersonaHub is a Hugging Face dataset focused on large-scale synthetic data creation using persona-driven generation. Its main purpose is to provide a massive collection of diverse personas that can be used to synthesize varied training data from different perspectives within an LLM. The dataset is text-based, provided ... | huggingface | proj-persona/PersonaHub |
SoloAI SFT Dataset: 20260425-2208
📊 数据集概览
| 属性 | 值 |
|---|---|
| 数据条数 | 30 条 |
| 数据源 | huggingface |
| 生成时间 | 2026-04-25 22:08 |
| 格式 | Instruction-Input-Output |
| 主要用途 | SFT 微调 / 指令调优 |
🎯 适用场景
- LLM 指令微调 (Instruction Tuning): 适用于训练对话型 AI 助手
- Prompt Engineering 研究: 可用于分析和优化提示词设计
- 多语言支持: 包含英文和中文数据
📝 数据格式
每条数据包含三个字段：
instruction: 任务指令（描述需要完成的任务）input: 输入上下文（提供任务相关的背景信息）output: 期望输出（模型应生成的回答）
示例
{'instruction': 'Summarize the dataset for LLM training use, including its purpose, likely use cases, scale, format, and notable tags.', 'input': 'Name: prompts.chat\nDescription: a.k.a. Awesome ChatGPT Prompts. This is a Dataset Repository mirror of prompts.chat — a social platform for AI prompts. prompts.chat is an open-source platform where users can share, discover, and collect AI prompts from the community.\nTags: task_categories:question-answering, task_categories:text-generation, license:cc0-1.0, size_categories:1K<n<10K, format:csv, modality:text, library:datasets, library:pandas, library:polars, library:mlcroissant, region:us, ChatGPT, prompts, AI, GPT, Claude, Gemini, Llama, Mistral, LLM, prompt-engineering, conversational-ai, text-generation, chatbot, awesome-list\nDownloads: 47783\nLikes: 9680', 'output': 'prompts.chat is a Hugging Face mirror of the Awesome ChatGPT Prompts community repository. It is a text dataset of shared AI prompts intended for prompt engineering, chatbot design, question answering, and text generation workflows. The dataset is relatively small, in the 1K to 10K size range, and is distributed in CSV format. Its CC0-1.0 license makes it easy to reuse. The tags indicate strong relevance to major LLM ecosystems such as ChatGPT, Claude, Gemini, Llama, and Mistral, making it useful as a prompt library, inspiration source, or seed dataset for instruction and conversational AI experiments.', 'source': 'huggingface', 'original_id': 'fka/prompts.chat'}
🤖 数据来源
本数据集由 SoloAI 自动化数据管道生成：
- 从 HuggingFace Datasets Hub 发现高质量数据集
- AI 清洗为 SFT 格式（Instruction-Input-Output）
- 质量过滤后发布
⚠️ 使用说明 & 📬 商务联系
- 本数据集仅供研究和实验用途
- 请遵守原始数据的许可证要求
- 商业用途 / 定制数据 / 深度合作:
- 📧 请联系:
379744050@qq.com - 🤖 SoloAI 提供高质量 SFT 数据定制服务。
- 建议邮件标题:
【数据定制咨询】行业 + 数据类型 + 规模 - 建议正文包含: 目标用途、需要条数、语言、字段格式、预算、交付时间
- 📧 请联系:
💰 商业合作报价
| 套餐 | 价格 | 说明 |
|---|---|---|
| Starter | $199 / 1000条高质量 SFT 数据 | 适合个人开发者 / 小团队 |
| Growth | $499 / 5000条行业数据 | 适合垂直行业训练数据 |
| Enterprise | $1499 / 定制领域数据管道 | 适合长期定制与数据管道 |
💳 支付方式
- 中国客户: 支付宝, 微信支付
- 海外客户: PayPal, USDT (TRC20)
- 下单方式: 邮件联系后 24 小时内提供交付方案与付款指引
🚀 为什么现在联系 SoloAI
- 24 小时内响应有效询盘
- 报价前可免费给出需求范围建议
- 支持中文 / English 项目合作
- 可从单次交付升级为长期数据管道合作
📈 更新日志
| 版本 | 日期 | 说明 |
|---|---|---|
| v1.0 | 2026-04-25 22:08 | 初始发布，30 条数据 |
- Downloads last month
- 22