instruction stringlengths 67 167 | input stringlengths 284 1.09k | output stringlengths 307 993 | source stringclasses 1
value | original_id stringlengths 8 44 |
|---|---|---|---|---|
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"fka/prompts.chat","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.... | prompts.chat is a mirrored Hugging Face dataset of community-shared AI prompts from prompts.chat, also associated with the Awesome ChatGPT Prompts project. It is a text dataset in CSV format, sized between 1K and 10K entries, and licensed under CC0-1.0, making it broadly reusable. The dataset is relevant for question a... | huggingface | fka/prompts.chat |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"HuggingFaceFW/fineweb","name":"fineweb","description":"FineWeb is a large-scale dataset of cleaned and deduplicated English web data from CommonCrawl, consisting of more than 18.5T tokens and processed for LLM performance using the datatrove library.","tags":["task_categories:text-generati... | FineWeb is a massive English web-text dataset derived from CommonCrawl, with over 18.5 trillion tokens of cleaned and deduplicated content. It is intended primarily for large-scale language model pretraining rather than direct supervised fine-tuning. The dataset is associated with text generation, includes both tabular... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"Anthropic/hh-rlhf","name":"hh-rlhf","description":"This repository provides 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... | HH-RLHF is a human-feedback dataset focused on helpfulness and harmlessness preferences, released by Anthropic. It is primarily designed for training preference or reward models in RLHF pipelines rather than for standard supervised fine-tuning of chat assistants. The dataset is text-based, provided in JSON format, lice... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"Open-Orca/OpenOrca","name":"OpenOrca","description":"The OpenOrca dataset is a rich collection of augmented FLAN data aligned as closely as possible with the distributions described in the Orca paper. It has been used to generate strong-performing model checkpoints.","tags":["task_categori... | OpenOrca is a large English instruction-style dataset built from augmented FLAN data and designed to approximate the data distribution described in the Orca work. It is well suited for supervised fine-tuning of general-purpose assistant models, especially for tasks such as question answering, summarization, text genera... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"OpenAssistant/oasst1","name":"oasst1","description":"OpenAssistant Conversations Dataset (OASST1) is a human-generated, human-annotated assistant-style conversation corpus with 161,443 messages in 35 languages, 461,292 quality ratings, and over 10,000 fully annotated conversation trees.","... | OASST1 is a multilingual assistant conversation dataset created through human generation and annotation, containing 161,443 messages across 35 languages, along with extensive quality ratings and annotated conversation trees. It is highly suitable for supervised fine-tuning of chat assistants, multilingual dialogue mode... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"openai/gsm8k","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 created to support question answering on basic mathematical problems that require multi-step reasoning. ... | GSM8K is a text dataset of English grade-school math word problems designed for multi-step reasoning and arithmetic question answering. It is well suited for SFT or evaluation of reasoning-capable models on math problem solving, especially chain-of-thought style tasks. The dataset is monolingual, original, crowdsourced... | huggingface | openai/gsm8k |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"wikimedia/wikipedia","name":"wikipedia","description":"Dataset Card for Wikimedia Wikipedia. Wikipedia dataset containing cleaned articles of all languages. The dataset is built from Wikipedia dumps with one subset per language, each containing a single train split. Each example contains t... | Wikimedia Wikipedia is a very large multilingual text corpus consisting of cleaned full Wikipedia articles across many languages. It is primarily useful for language modeling, masked language modeling, pretraining, retrieval, and knowledge-rich text generation rather than narrow instruction tuning. Its notable properti... | huggingface | wikimedia/wikipedia |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"gsdf/EasyNegative","name":"EasyNegative","description":"Negative Embedding trained with Counterfeit. Intended for use in the stable-diffusion-webui embeddings folder and may be usable with other models, though effectiveness is uncertain.","tags":["license:other","size_categories:n<1K","for... | EasyNegative is an image-related asset centered on a negative embedding for Stable Diffusion workflows, not a conventional text instruction dataset. It is relevant to image generation and prompt conditioning rather than LLM SFT. The item is very small, uses imagefolder format, and has a nonstandard 'other' license. For... | huggingface | gsdf/EasyNegative |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"togethercomputer/RedPajama-Data-1T","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:ml... | RedPajama-Data-1T is a large-scale English text corpus intended for text generation and open LLM pretraining. It is best suited for foundation-model pretraining, continued pretraining, and large-scale corpus mining rather than direct supervised instruction tuning. A key property is that it is positioned as a clean-room... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"FreedomIntelligence/medical-o1-reasoning-SFT","name":"medical-o1-reasoning-SFT","description":"Medical reasoning dataset for SFT built on medical verifiable problems and LLM-generated reasoning chains. The card notes releases of distilled data from Deepseek-R1 and a split keeping only the ... | medical-o1-reasoning-SFT is a bilingual English-Chinese medical text dataset explicitly designed for supervised fine-tuning on reasoning-heavy medical question answering and generation. It is highly relevant for domain-specific SFT because it focuses on verifiable medical problems and includes reasoning-oriented distil... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset for language model pretraining use, including scale, domain, language, license, and notable characteristics. | Name: fineweb-edu
Description: FineWeb-Edu is a dataset of 1.3 trillion tokens of educational web data filtered from FineWeb using an educational quality classifier. It is the 1.3T token version and is intended for high-quality educational language modeling.
Tags: task_categories:text-generation, language:en, license:o... | FineWeb-Edu is a large-scale English educational web corpus for text generation and language model pretraining. It contains about 1.3 trillion tokens filtered from the FineWeb dataset using an educational quality classifier, making it especially useful for training models on high-quality educational content. The datase... | huggingface | HuggingFaceFW/fineweb-edu |
Provide a concise dataset summary focused on pretraining relevance, corpus size, language, and licensing. | Name: dolma
Description: Dolma is 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, language-modeling, casual-lm, llm
Downloads: 3153
Likes: 1023 | Dolma is an open English corpus designed for language model pretraining research. It contains roughly three trillion tokens, making it suitable for large-scale LLM training and language modeling experiments. The dataset is tagged for text generation and LLM use, and it is released under the ODC-BY license. | huggingface | allenai/dolma |
Summarize the dataset with emphasis on content type, scale, multilingual/code coverage, and licensing considerations. | Name: the-stack
Description: The Stack is a large code dataset. Initial release included 30 programming languages and 18 permissive licenses, producing a near-deduplicated 3TB dataset. Later versions excluded weak copyleft licenses and expanded permissive license coverage.
Tags: task_categories:text-generation, multili... | The Stack is a multilingual code dataset built for text generation and code language modeling. It contains source code across many programming languages and was released as a large near-deduplicated corpus measuring multiple terabytes. A key feature is its focus on license filtering, with later releases excluding weak ... | huggingface | bigcode/the-stack |
Create a short dataset description covering data origin, task fit, language, size, and special properties. | Name: TinyStories
Description: TinyStories contains synthetically generated short stories created by GPT-3.5 and GPT-4 using a small vocabulary. It is associated with the paper at arXiv:2305.07759 and is used for training compact language models.
Tags: task_categories:text-generation, language:en, license:cdla-sharing-... | TinyStories is an English text generation dataset composed of synthetic short stories generated by GPT-3.5 and GPT-4. Its stories intentionally use a limited vocabulary, making it well suited for studying language acquisition, small-model training, and controlled language modeling experiments. The dataset is relatively... | huggingface | roneneldan/TinyStories |
Summarize the dataset as an instruction-tuning resource, noting task types, scale, language, and license. | Name: databricks-dolly-15k
Description: databricks-dolly-15k is an open source instruction-following dataset created by Databricks employees across categories such as brainstorming, classification, closed QA, generation, information extraction, open QA, and summarization.
Tags: task_categories:question-answering, task_... | databricks-dolly-15k is an English instruction-tuning dataset containing about 15,000 human-generated records for tasks such as brainstorming, classification, question answering, information extraction, generation, and summarization. It is intended for training instruction-following models and is a popular open resourc... | huggingface | databricks/databricks-dolly-15k |
Summarize the following Hugging Face dataset entry for supervised fine-tuning. | {"source":"huggingface","id":"tatsu-lab/alpaca","name":"alpaca","description":"Dataset Card for Alpaca\n\nDataset Summary\n\nAlpaca 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 an... | Alpaca is an English text dataset on Hugging Face designed for instruction fine-tuning. It contains about 52,000 instruction-response demonstrations generated with OpenAI's text-davinci-003 and built on the Self-Instruct pipeline. The dataset is suitable for text generation and instruction-following model training. It ... | huggingface | tatsu-lab/alpaca |
Summarize the following Hugging Face dataset entry for supervised fine-tuning. | {"source":"huggingface","id":"Nerfgun3/bad_prompt","name":"bad_prompt","description":"Negative Embedding / Textual Inversion\n\nIdea\n\nThe 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. Side note: Embedding... | bad_prompt is a small English image-related Hugging Face dataset/resource associated with Stable Diffusion and textual inversion. It focuses on a negative embedding concept intended to consolidate negative prompting into a reusable embedding, with noted usefulness for image generation quality such as hands. The dataset... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset and list its key metadata for supervised fine-tuning. | Name: falcon-refinedweb
Source: huggingface
Dataset ID: tiiuae/falcon-refinedweb
Description: Falcon RefinedWeb is a massive English web dataset built by TII and released under an ODC-By 1.0 license. It is created through stringent filtering and large-scale deduplication of CommonCrawl. The dataset is intended for text... | Falcon RefinedWeb is a large English web corpus from TII designed for text generation research and model training. It is built from CommonCrawl using strong filtering and large-scale deduplication, and is positioned as a high-quality web-only alternative to curated datasets. The dataset is licensed under ODC-By 1.0, st... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset and list its key metadata for supervised fine-tuning. | Name: lmsys-chat-1m
Source: huggingface
Dataset ID: lmsys/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 between ... | LMSYS-Chat-1M is a real-world conversation dataset containing one million chats involving 25 leading LLMs. The data was collected from the Vicuna demo and Chatbot Arena platforms from April to August 2023, covering activity from 210K unique IP addresses. Each record includes a conversation ID, model name, conversation ... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise SFT-style description including its purpose, source, notable characteristics, and basic popularity signals. | Name: ShareGPT_Vicuna_unfiltered
Source: huggingface
Dataset ID: anon8231489123/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 are mentioned: a version removing in... | ShareGPT_Vicuna_unfiltered is a Hugging Face dataset intended for conversational model training and refinement. It appears to be a cleaned ShareGPT/Vicuna-style corpus, with notes about additional cleaning and verified training usage. The description highlights two dataset variants, including one that removes occurrenc... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise SFT-style description including its purpose, source, notable characteristics, and basic popularity signals. | Name: PhysicalAI-Autonomous-Vehicles
Source: huggingface
Dataset ID: nvidia/PhysicalAI-Autonomous-Vehicles
Description: PHYSICAL AI AUTONOMOUS VEHICLES. The dataset provides one of the largest geographically diverse collections of multi-sensor data for AV researchers to build next-generation Physical AI end-to-end driv... | PhysicalAI-Autonomous-Vehicles is a Hugging Face dataset from NVIDIA for autonomous vehicle and physical AI research. It offers a large, geographically diverse multi-sensor driving dataset designed to support end-to-end driving system development. The description states that it is licensed for both commercial and non-c... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset based on the provided metadata, including what it contains, scale, language coverage, format, licensing, and notable usage details. | {"source":"huggingface","id":"HuggingFaceFW/finepdfs","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_categori... | finepdfs is a large-scale Hugging Face dataset built exclusively from PDF sources. It contains approximately 3 trillion tokens extracted from 475 million documents and covers 1,733 languages, making it a major multilingual corpus for text-generation and language model training. The dataset is distributed in Parquet for... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset based on the provided metadata, including domain, scale, data type, intended use, format, licensing, and notable characteristics. | {"source":"huggingface","id":"open-thoughts/OpenThoughts-114k","name":"OpenThoughts-114k","description":"Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. Default subset contains ready-to-train data used to finetune the OpenThinker-7B and OpenThinker-32B models.... | OpenThoughts-114k is a synthetic reasoning dataset hosted on Hugging Face. It provides 114,000 high-quality text examples spanning math, science, code, and puzzle-solving tasks, and is designed for model fine-tuning. The default subset contains ready-to-train data that was used to finetune the OpenThinker-7B and OpenTh... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction-tuning use, including its purpose, scale, format, and notable characteristics. | 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. The Open Hermes 2/2.5 and Nous Hermes 2 models have made significant advancements of SOTA LLM's over recent months, and are underpinned by this exact ... | OpenHermes-2.5 is a large English text dataset on Hugging Face designed for LLM training and instruction tuning. It is the dataset behind OpenHermes 2.5 and Nous Hermes 2 model series, and consists of a curated compilation of open-source datasets along with custom synthetic data. The dataset is provided in JSON format,... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction-tuning use, including its purpose, scale, format, and notable characteristics. | Name: alpaca-cleaned
Source: huggingface
Dataset ID: yahma/alpaca-cleaned
Description: This is a cleaned version of the original Alpaca Dataset released by Stanford. The following issues have been identified in the original release and fixed in this dataset: Hallucinations: Many instructions in the original dataset had... | alpaca-cleaned is an English instruction-finetuning dataset on Hugging Face derived from Stanford's original Alpaca release. It was created to address quality issues in the original dataset, especially examples that encouraged hallucinated responses by referencing unavailable internet information. The dataset is intend... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset metadata into a concise SFT training example. | Source: huggingface
Dataset 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. The datase... | {"name":"hle","provider":"cais","full_id":"cais/hle","summary":"Humanity's Last Exam is a multimodal benchmark dataset containing 2,500 closed-ended questions spanning many academic subjects at the frontier of human knowledge.","modalities":["image","text"],"format":"parquet","license":"mit","size_category":"1K<n<10K",... | huggingface | cais/hle |
Summarize the dataset metadata into a concise SFT training example. | Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-2
Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining text data for over 1000 languages. It is fully reproducible, extensively validated through hundreds of ablation experiments, and released under... | {"name":"fineweb-2","provider":"HuggingFaceFW","full_id":"HuggingFaceFW/fineweb-2","summary":"FineWeb2 is a large-scale multilingual text pretraining dataset covering over 1000 languages, designed for high-quality and reproducible language model training.","task_category":"text-generation","modalities":["tabular","text... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset based on the provided metadata, including its purpose, modality, likely tasks, scale, language, license, popularity, and notable characteristics. | {"source":"huggingface","id":"ILSVRC/imagenet-1k","name":"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 synset. ImageNet aims to... | ImageNet-1k is a large-scale original image dataset from Hugging Face based on ILSVRC 2012. It is designed primarily for multi-class image classification and contains images organized by WordNet synsets, with roughly 1,000 images per concept on average. The dataset modality is image, the language metadata indicates Eng... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset based on the provided metadata, including its purpose, data type, languages, license, popularity, and notable characteristics. | {"source":"huggingface","id":"QingyiSi/Alpaca-CoT","name":"Alpaca-CoT","description":"This repository continuously collects various instruction tuning datasets and standardizes them into the same format so they can be directly loaded by Alpaca model code. It also accompanies empirical study on various instruction-tunin... | Alpaca-CoT is an instruction-finetuning dataset collection hosted on Hugging Face. Its purpose is to aggregate various instruction-tuning datasets and standardize them into a unified format that can be directly used with Alpaca-style model training pipelines. The tags suggest it is multilingual, including English, Chin... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset PersonaHub, including its purpose, scale, modalities, languages, tasks, license, and notable metadata. | Name: PersonaHub
Source: huggingface
ID: proj-persona/PersonaHub
Description: Scaling Synthetic Data Creation with 1,000,000,000 Personas. This repository 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 lev... | PersonaHub is a Hugging Face dataset focused on scalable synthetic data creation through persona-driven generation. Introduced in the paper 'Scaling Synthetic Data Creation with 1,000,000,000 Personas', it presents PERSONA HUB, a large collection of 1 billion diverse personas automatically curated from web data to help... | huggingface | proj-persona/PersonaHub |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"fka/prompts.chat","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.... | prompts.chat is a mirrored Hugging Face dataset of community-shared AI prompts from prompts.chat, also associated with the Awesome ChatGPT Prompts project. It is a text dataset in CSV format, sized between 1K and 10K entries, and licensed under CC0-1.0, making it broadly reusable. The dataset is relevant for question a... | huggingface | fka/prompts.chat |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"HuggingFaceFW/fineweb","name":"fineweb","description":"FineWeb is a large-scale dataset of cleaned and deduplicated English web data from CommonCrawl, consisting of more than 18.5T tokens and processed for LLM performance using the datatrove library.","tags":["task_categories:text-generati... | FineWeb is a massive English web-text dataset derived from CommonCrawl, with over 18.5 trillion tokens of cleaned and deduplicated content. It is intended primarily for large-scale language model pretraining rather than direct supervised fine-tuning. The dataset is associated with text generation, includes both tabular... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"Anthropic/hh-rlhf","name":"hh-rlhf","description":"This repository provides 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... | HH-RLHF is a human-feedback dataset focused on helpfulness and harmlessness preferences, released by Anthropic. It is primarily designed for training preference or reward models in RLHF pipelines rather than for standard supervised fine-tuning of chat assistants. The dataset is text-based, provided in JSON format, lice... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"Open-Orca/OpenOrca","name":"OpenOrca","description":"The OpenOrca dataset is a rich collection of augmented FLAN data aligned as closely as possible with the distributions described in the Orca paper. It has been used to generate strong-performing model checkpoints.","tags":["task_categori... | OpenOrca is a large English instruction-style dataset built from augmented FLAN data and designed to approximate the data distribution described in the Orca work. It is well suited for supervised fine-tuning of general-purpose assistant models, especially for tasks such as question answering, summarization, text genera... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for LLM supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions. | {"source":"huggingface","id":"OpenAssistant/oasst1","name":"oasst1","description":"OpenAssistant Conversations Dataset (OASST1) is a human-generated, human-annotated assistant-style conversation corpus with 161,443 messages in 35 languages, 461,292 quality ratings, and over 10,000 fully annotated conversation trees.","... | OASST1 is a multilingual assistant conversation dataset created through human generation and annotation, containing 161,443 messages across 35 languages, along with extensive quality ratings and annotated conversation trees. It is highly suitable for supervised fine-tuning of chat assistants, multilingual dialogue mode... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"openai/gsm8k","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 created to support question answering on basic mathematical problems that require multi-step reasoning. ... | GSM8K is a text dataset of English grade-school math word problems designed for multi-step reasoning and arithmetic question answering. It is well suited for SFT or evaluation of reasoning-capable models on math problem solving, especially chain-of-thought style tasks. The dataset is monolingual, original, crowdsourced... | huggingface | openai/gsm8k |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"wikimedia/wikipedia","name":"wikipedia","description":"Dataset Card for Wikimedia Wikipedia. Wikipedia dataset containing cleaned articles of all languages. The dataset is built from Wikipedia dumps with one subset per language, each containing a single train split. Each example contains t... | Wikimedia Wikipedia is a very large multilingual text corpus consisting of cleaned full Wikipedia articles across many languages. It is primarily useful for language modeling, masked language modeling, pretraining, retrieval, and knowledge-rich text generation rather than narrow instruction tuning. Its notable properti... | huggingface | wikimedia/wikipedia |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"gsdf/EasyNegative","name":"EasyNegative","description":"Negative Embedding trained with Counterfeit. Intended for use in the stable-diffusion-webui embeddings folder and may be usable with other models, though effectiveness is uncertain.","tags":["license:other","size_categories:n<1K","for... | EasyNegative is an image-related asset centered on a negative embedding for Stable Diffusion workflows, not a conventional text instruction dataset. It is relevant to image generation and prompt conditioning rather than LLM SFT. The item is very small, uses imagefolder format, and has a nonstandard 'other' license. For... | huggingface | gsdf/EasyNegative |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"togethercomputer/RedPajama-Data-1T","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:ml... | RedPajama-Data-1T is a large-scale English text corpus intended for text generation and open LLM pretraining. It is best suited for foundation-model pretraining, continued pretraining, and large-scale corpus mining rather than direct supervised instruction tuning. A key property is that it is positioned as a clean-room... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset for supervised fine-tuning relevance, including domain, modality, likely use cases, and notable properties. | {"source":"huggingface","id":"FreedomIntelligence/medical-o1-reasoning-SFT","name":"medical-o1-reasoning-SFT","description":"Medical reasoning dataset for SFT built on medical verifiable problems and LLM-generated reasoning chains. The card notes releases of distilled data from Deepseek-R1 and a split keeping only the ... | medical-o1-reasoning-SFT is a bilingual English-Chinese medical text dataset explicitly designed for supervised fine-tuning on reasoning-heavy medical question answering and generation. It is highly relevant for domain-specific SFT because it focuses on verifiable medical problems and includes reasoning-oriented distil... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset for language model pretraining use, including scale, domain, language, license, and notable characteristics. | Name: fineweb-edu
Description: FineWeb-Edu is a dataset of 1.3 trillion tokens of educational web data filtered from FineWeb using an educational quality classifier. It is the 1.3T token version and is intended for high-quality educational language modeling.
Tags: task_categories:text-generation, language:en, license:o... | FineWeb-Edu is a large-scale English educational web corpus for text generation and language model pretraining. It contains about 1.3 trillion tokens filtered from the FineWeb dataset using an educational quality classifier, making it especially useful for training models on high-quality educational content. The datase... | huggingface | HuggingFaceFW/fineweb-edu |
Provide a concise dataset summary focused on pretraining relevance, corpus size, language, and licensing. | Name: dolma
Description: Dolma is 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, language-modeling, casual-lm, llm
Downloads: 3153
Likes: 1023 | Dolma is an open English corpus designed for language model pretraining research. It contains roughly three trillion tokens, making it suitable for large-scale LLM training and language modeling experiments. The dataset is tagged for text generation and LLM use, and it is released under the ODC-BY license. | huggingface | allenai/dolma |
Summarize the dataset with emphasis on content type, scale, multilingual/code coverage, and licensing considerations. | Name: the-stack
Description: The Stack is a large code dataset. Initial release included 30 programming languages and 18 permissive licenses, producing a near-deduplicated 3TB dataset. Later versions excluded weak copyleft licenses and expanded permissive license coverage.
Tags: task_categories:text-generation, multili... | The Stack is a multilingual code dataset built for text generation and code language modeling. It contains source code across many programming languages and was released as a large near-deduplicated corpus measuring multiple terabytes. A key feature is its focus on license filtering, with later releases excluding weak ... | huggingface | bigcode/the-stack |
Create a short dataset description covering data origin, task fit, language, size, and special properties. | Name: TinyStories
Description: TinyStories contains synthetically generated short stories created by GPT-3.5 and GPT-4 using a small vocabulary. It is associated with the paper at arXiv:2305.07759 and is used for training compact language models.
Tags: task_categories:text-generation, language:en, license:cdla-sharing-... | TinyStories is an English text generation dataset composed of synthetic short stories generated by GPT-3.5 and GPT-4. Its stories intentionally use a limited vocabulary, making it well suited for studying language acquisition, small-model training, and controlled language modeling experiments. The dataset is relatively... | huggingface | roneneldan/TinyStories |
Summarize the dataset as an instruction-tuning resource, noting task types, scale, language, and license. | Name: databricks-dolly-15k
Description: databricks-dolly-15k is an open source instruction-following dataset created by Databricks employees across categories such as brainstorming, classification, closed QA, generation, information extraction, open QA, and summarization.
Tags: task_categories:question-answering, task_... | databricks-dolly-15k is an English instruction-tuning dataset containing about 15,000 human-generated records for tasks such as brainstorming, classification, question answering, information extraction, generation, and summarization. It is intended for training instruction-following models and is a popular open resourc... | huggingface | databricks/databricks-dolly-15k |
Summarize the following Hugging Face dataset entry for supervised fine-tuning. | {"source":"huggingface","id":"tatsu-lab/alpaca","name":"alpaca","description":"Dataset Card for Alpaca\n\nDataset Summary\n\nAlpaca 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 an... | Alpaca is an English text dataset on Hugging Face designed for instruction fine-tuning. It contains about 52,000 instruction-response demonstrations generated with OpenAI's text-davinci-003 and built on the Self-Instruct pipeline. The dataset is suitable for text generation and instruction-following model training. It ... | huggingface | tatsu-lab/alpaca |
Summarize the following Hugging Face dataset entry for supervised fine-tuning. | {"source":"huggingface","id":"Nerfgun3/bad_prompt","name":"bad_prompt","description":"Negative Embedding / Textual Inversion\n\nIdea\n\nThe 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. Side note: Embedding... | bad_prompt is a small English image-related Hugging Face dataset/resource associated with Stable Diffusion and textual inversion. It focuses on a negative embedding concept intended to consolidate negative prompting into a reusable embedding, with noted usefulness for image generation quality such as hands. The dataset... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset and list its key metadata for supervised fine-tuning. | Name: falcon-refinedweb
Source: huggingface
Dataset ID: tiiuae/falcon-refinedweb
Description: Falcon RefinedWeb is a massive English web dataset built by TII and released under an ODC-By 1.0 license. It is created through stringent filtering and large-scale deduplication of CommonCrawl. The dataset is intended for text... | Falcon RefinedWeb is a large English web corpus from TII designed for text generation research and model training. It is built from CommonCrawl using strong filtering and large-scale deduplication, and is positioned as a high-quality web-only alternative to curated datasets. The dataset is licensed under ODC-By 1.0, st... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset and list its key metadata for supervised fine-tuning. | Name: lmsys-chat-1m
Source: huggingface
Dataset ID: lmsys/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 between ... | LMSYS-Chat-1M is a real-world conversation dataset containing one million chats involving 25 leading LLMs. The data was collected from the Vicuna demo and Chatbot Arena platforms from April to August 2023, covering activity from 210K unique IP addresses. Each record includes a conversation ID, model name, conversation ... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise SFT-style description including its purpose, source, notable characteristics, and basic popularity signals. | Name: ShareGPT_Vicuna_unfiltered
Source: huggingface
Dataset ID: anon8231489123/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 are mentioned: a version removing in... | ShareGPT_Vicuna_unfiltered is a Hugging Face dataset intended for conversational model training and refinement. It appears to be a cleaned ShareGPT/Vicuna-style corpus, with notes about additional cleaning and verified training usage. The description highlights two dataset variants, including one that removes occurrenc... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise SFT-style description including its purpose, source, notable characteristics, and basic popularity signals. | Name: PhysicalAI-Autonomous-Vehicles
Source: huggingface
Dataset ID: nvidia/PhysicalAI-Autonomous-Vehicles
Description: PHYSICAL AI AUTONOMOUS VEHICLES. The dataset provides one of the largest geographically diverse collections of multi-sensor data for AV researchers to build next-generation Physical AI end-to-end driv... | PhysicalAI-Autonomous-Vehicles is a Hugging Face dataset from NVIDIA for autonomous vehicle and physical AI research. It offers a large, geographically diverse multi-sensor driving dataset designed to support end-to-end driving system development. The description states that it is licensed for both commercial and non-c... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset based on the provided metadata, including what it contains, scale, language coverage, format, licensing, and notable usage details. | {"source":"huggingface","id":"HuggingFaceFW/finepdfs","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_categori... | finepdfs is a large-scale Hugging Face dataset built exclusively from PDF sources. It contains approximately 3 trillion tokens extracted from 475 million documents and covers 1,733 languages, making it a major multilingual corpus for text-generation and language model training. The dataset is distributed in Parquet for... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset based on the provided metadata, including domain, scale, data type, intended use, format, licensing, and notable characteristics. | {"source":"huggingface","id":"open-thoughts/OpenThoughts-114k","name":"OpenThoughts-114k","description":"Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. Default subset contains ready-to-train data used to finetune the OpenThinker-7B and OpenThinker-32B models.... | OpenThoughts-114k is a synthetic reasoning dataset hosted on Hugging Face. It provides 114,000 high-quality text examples spanning math, science, code, and puzzle-solving tasks, and is designed for model fine-tuning. The default subset contains ready-to-train data that was used to finetune the OpenThinker-7B and OpenTh... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction-tuning use, including its purpose, scale, format, and notable characteristics. | 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. The Open Hermes 2/2.5 and Nous Hermes 2 models have made significant advancements of SOTA LLM's over recent months, and are underpinned by this exact ... | OpenHermes-2.5 is a large English text dataset on Hugging Face designed for LLM training and instruction tuning. It is the dataset behind OpenHermes 2.5 and Nous Hermes 2 model series, and consists of a curated compilation of open-source datasets along with custom synthetic data. The dataset is provided in JSON format,... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction-tuning use, including its purpose, scale, format, and notable characteristics. | Name: alpaca-cleaned
Source: huggingface
Dataset ID: yahma/alpaca-cleaned
Description: This is a cleaned version of the original Alpaca Dataset released by Stanford. The following issues have been identified in the original release and fixed in this dataset: Hallucinations: Many instructions in the original dataset had... | alpaca-cleaned is an English instruction-finetuning dataset on Hugging Face derived from Stanford's original Alpaca release. It was created to address quality issues in the original dataset, especially examples that encouraged hallucinated responses by referencing unavailable internet information. The dataset is intend... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset metadata into a concise SFT training example. | Source: huggingface
Dataset 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. The datase... | {"name":"hle","provider":"cais","full_id":"cais/hle","summary":"Humanity's Last Exam is a multimodal benchmark dataset containing 2,500 closed-ended questions spanning many academic subjects at the frontier of human knowledge.","modalities":["image","text"],"format":"parquet","license":"mit","size_category":"1K<n<10K",... | huggingface | cais/hle |
Summarize the dataset metadata into a concise SFT training example. | Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-2
Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining text data for over 1000 languages. It is fully reproducible, extensively validated through hundreds of ablation experiments, and released under... | {"name":"fineweb-2","provider":"HuggingFaceFW","full_id":"HuggingFaceFW/fineweb-2","summary":"FineWeb2 is a large-scale multilingual text pretraining dataset covering over 1000 languages, designed for high-quality and reproducible language model training.","task_category":"text-generation","modalities":["tabular","text... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset based on the provided metadata, including its purpose, modality, likely tasks, scale, language, license, popularity, and notable characteristics. | {"source":"huggingface","id":"ILSVRC/imagenet-1k","name":"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 synset. ImageNet aims to... | ImageNet-1k is a large-scale original image dataset from Hugging Face based on ILSVRC 2012. It is designed primarily for multi-class image classification and contains images organized by WordNet synsets, with roughly 1,000 images per concept on average. The dataset modality is image, the language metadata indicates Eng... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset based on the provided metadata, including its purpose, data type, languages, license, popularity, and notable characteristics. | {"source":"huggingface","id":"QingyiSi/Alpaca-CoT","name":"Alpaca-CoT","description":"This repository continuously collects various instruction tuning datasets and standardizes them into the same format so they can be directly loaded by Alpaca model code. It also accompanies empirical study on various instruction-tunin... | Alpaca-CoT is an instruction-finetuning dataset collection hosted on Hugging Face. Its purpose is to aggregate various instruction-tuning datasets and standardize them into a unified format that can be directly used with Alpaca-style model training pipelines. The tags suggest it is multilingual, including English, Chin... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset PersonaHub, including its purpose, scale, modalities, languages, tasks, license, and notable metadata. | Name: PersonaHub
Source: huggingface
ID: proj-persona/PersonaHub
Description: Scaling Synthetic Data Creation with 1,000,000,000 Personas. This repository 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 lev... | PersonaHub is a Hugging Face dataset focused on scalable synthetic data creation through persona-driven generation. Introduced in the paper 'Scaling Synthetic Data Creation with 1,000,000,000 Personas', it presents PERSONA HUB, a large collection of 1 billion diverse personas automatically curated from web data to help... | huggingface | proj-persona/PersonaHub |
SoloAI SFT Dataset: 20260428-0940
📊 数据集概览
| 属性 | 值 |
|---|---|
| 数据条数 | 30 条 |
| 数据源 | huggingface |
| 生成时间 | 2026-04-28 09:40 |
| 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 supervised fine-tuning use, including what it contains, likely use cases, notable metadata, and any cautions.', 'input': '{"source":"huggingface","id":"fka/prompts.chat","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_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"],"downloads":53719,"likes":9680,"created_at":""}', 'output': 'prompts.chat is a mirrored Hugging Face dataset of community-shared AI prompts from prompts.chat, also associated with the Awesome ChatGPT Prompts project. It is a text dataset in CSV format, sized between 1K and 10K entries, and licensed under CC0-1.0, making it broadly reusable. The dataset is relevant for question answering and text generation workflows, especially prompt engineering research, prompt library construction, chatbot prototyping, and comparative prompting across models such as ChatGPT, Claude, Gemini, Llama, and Mistral. A key strength is its popularity and practical focus on real-world prompts. A caution is that this is primarily a prompt collection rather than a curated instruction-response corpus, so it may need preprocessing or augmentation before direct SFT use.', '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 | 2026-04-28 09:40 | 初始发布，30 条数据 |
- Downloads last month
- 21