instruction stringlengths 53 217 | input stringlengths 349 1.04k | output stringlengths 357 853 | source stringclasses 1
value | original_id stringlengths 8 44 |
|---|---|---|---|---|
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"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 small open prompt collection mirrored from the prompts.chat community platform. It is best suited for prompt engineering research, prompt retrieval, instruction template discovery, chatbot ideation, and benchmarking prompt diversity across popular LLM ecosystems such as ChatGPT, Claude, Gemini, Llama,... | huggingface | fka/prompts.chat |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"HuggingFaceFW/fineweb","name":"fineweb","description":"FineWeb: 15 trillion tokens of the finest data the web has to offer. The FineWeb dataset consists of more than 18.5T tokens of cleaned and deduplicated English web data from CommonCrawl. The data processing pipeline is optimized for LL... | FineWeb is a massive English web-scale corpus built from cleaned and deduplicated CommonCrawl data, intended primarily for large language model pretraining. Its defining characteristic is scale: over 18.5 trillion tokens, making it far larger than typical supervised fine-tuning datasets. It is useful for continued pret... | huggingface | HuggingFaceFW/fineweb |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"Anthropic/hh-rlhf","name":"hh-rlhf","description":"Dataset Card for HH-RLHF. 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 t... | HH-RLHF is a human feedback dataset focused on helpfulness and harmlessness preferences, designed for training reward models or preference models in RLHF pipelines. It is a medium-to-large text dataset in JSON format with roughly 100K to 1M examples and an MIT license. Its strongest use cases are preference learning, r... | huggingface | Anthropic/hh-rlhf |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"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 outlined in the Orca paper. It has been instrumental in generating high-performing model checkpoints and serves as a v... | OpenOrca is a large English instruction-style dataset built from augmented FLAN-style data and inspired by the Orca paper’s distribution. It is highly relevant for supervised fine-tuning because it covers a broad range of NLP tasks, including question answering, summarization, classification, feature extraction, and te... | huggingface | Open-Orca/OpenOrca |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"OpenAssistant/oasst1","name":"oasst1","description":"OpenAssistant Conversations Dataset (OASST1). A human-generated, human-annotated assistant-style conversation corpus consisting of 161,443 messages in 35 different languages, annotated with 461,292 quality ratings, resulting in over 10,0... | OASST1 is a multilingual, human-generated and human-annotated assistant conversation dataset created for alignment research. It contains over 161K messages across 35 languages, along with quality ratings and conversation tree structure, making it especially valuable for supervised fine-tuning of chat assistants, multil... | huggingface | OpenAssistant/oasst1 |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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. The dataset was created to support the task of question answering on basic mathematical problems that r... | gsm8k is a Hugging Face dataset focused on grade school math word problems that require multi-step reasoning. It is intended for question answering and text-generation style evaluation on basic arithmetic reasoning. The data is text-only, monolingual English, and tagged as an official benchmark with original crowdsourc... | huggingface | openai/gsm8k |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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 the Wikipedia dumps with one subset per language, each containing a single train split. Each example contai... | wikipedia is a Hugging Face dataset containing cleaned full Wikipedia articles across many languages, derived from Wikimedia dumps. It is primarily useful for language modeling, masked language modeling, text generation, and fill-mask tasks. The modality is text, and the dataset is massively multilingual with one subse... | huggingface | wikimedia/wikipedia |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"source":"huggingface","id":"gsdf/EasyNegative","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"... | EasyNegative is a Hugging Face image-related asset described as a negative embedding trained with Counterfeit for use with Stable Diffusion WebUI embeddings. Its modality is image, and it appears intended for generative image workflows rather than conventional NLP tasks. Notable metadata includes imagefolder format, ve... | huggingface | gsdf/EasyNegative |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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 Hugging Face text dataset positioned as a clean-room, fully open-source implementation of the LLaMa dataset. Its main purpose is text generation and large-scale language model training, and the metadata indicates English text. Notable tags show text modality, large size, and compatibility with da... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"source":"huggingface","id":"FreedomIntelligence/medical-o1-reasoning-SFT","name":"medical-o1-reasoning-SFT","description":"News: [2025/04/22] We split the data and kept only the medical SFT dataset (medical_o1_sft.json). The file medical_o1_sft_mix.json contains a mix of medical and general instruction data. [2025/02... | medical-o1-reasoning-SFT is a Hugging Face dataset for medical reasoning supervised fine-tuning, built around medical verifiable problems and distilled reasoning data. It supports question answering and text generation tasks in text modality, with English and Chinese language coverage. Notable metadata includes Apache-... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"HuggingFaceFW/fineweb-edu","name":"fineweb-edu","description":"1.3 trillion tokens of the finest educational data the web has to offer. FineWeb-Edu consists of 1.3T tokens of educational web pages filtered from FineWeb using an educational quality classifier with annotations generated by L... | fineweb-edu is a Hugging Face dataset for text generation focused on high-quality educational web content in English. It contains about 1.3 trillion tokens filtered from FineWeb using an educational quality classifier, making it useful for large-scale language model pretraining. The dataset is available in parquet form... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"allenai/dolma","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-m... | dolma is a Hugging Face dataset for text generation and language model pretraining research. It is an open English corpus containing roughly three trillion tokens, designed for large-scale LLM training and language modeling use cases. The dataset is licensed under ODC-By and references papers such as arXiv:2402.00159. ... | huggingface | allenai/dolma |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"bigcode/the-stack","name":"the-stack","description":"Dataset Card for The Stack. Initial release included 30 programming languages and 18 permissive licenses, with a near-deduplicated dataset size of 3TB. Later versions excluded weak copyleft licenses and expanded the permissive license li... | the-stack is a Hugging Face dataset for text generation over source code. It is a multilingual code corpus built from crowdsourced and expert-generated repositories, near-deduplicated, and originally around 3TB in size. The dataset is distributed in parquet format with text and tabular modalities, uses a nonstandard or... | huggingface | bigcode/the-stack |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"roneneldan/TinyStories","name":"TinyStories","description":"Dataset containing synthetically generated by GPT-3.5 and GPT-4 short stories that only use a small vocabulary. Described in arXiv:2305.07759 and used to train small language models.","tags":["task_categories:text-generation","lan... | TinyStories is a Hugging Face dataset for English text generation composed of synthetic short stories generated by GPT-3.5 and GPT-4 using a restricted vocabulary. It is intended for training and studying small language models, with a dataset size in the 1M to 10M range. The dataset is available in parquet format, cont... | huggingface | roneneldan/TinyStories |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"databricks/databricks-dolly-15k","name":"databricks-dolly-15k","description":"databricks-dolly-15k is an open source dataset of instruction-following records generated by thousands of Databricks employees across categories including brainstorming, classification, closed QA, generation, inf... | databricks-dolly-15k is a Hugging Face instruction-tuning dataset in English containing open-source instruction-following examples created by Databricks employees. It covers tasks such as question answering, summarization, brainstorming, classification, generation, and information extraction, with a size between 10K an... | huggingface | databricks/databricks-dolly-15k |
Summarize the given dataset entry for instruction-tuning relevance, modality, language, license, scale, and notable metadata. | Name: alpaca
Source: huggingface
ID: tatsu-lab/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. The a... | Alpaca is an English text dataset on Hugging Face designed for instruction fine-tuning. It contains about 52,000 instruction-response demonstrations generated with text-davinci-003 and based on the Self-Instruct pipeline. The dataset is tagged for text generation, uses Parquet format, and falls in the 10K to 100K size ... | huggingface | tatsu-lab/alpaca |
Summarize the given dataset entry, highlighting what it is for, modality, usage context, license, size, and key tags. | Name: bad_prompt
Source: huggingface
ID: Nerfgun3/bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to train the negative prompt as an embedding, unifying the basis of the negative prompt into one word or embedding. It is noted as helpful for image generation, especially... | bad_prompt is a small English Hugging Face image-related resource associated with Stable Diffusion workflows. It is described as a negative embedding or textual inversion asset intended to compress negative prompt behavior into a reusable embedding, with claimed benefits for generation quality such as improved hands. T... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset metadata into a concise SFT training example. | 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 built through stringent filtering and large-scale deduplication of CommonCrawl. The dataset is described as achiev... | Falcon RefinedWeb is a large English web text dataset from TII on Hugging Face. It is created from CommonCrawl using strict filtering and large-scale deduplication, and is licensed under ODC-By 1.0. The dataset targets text generation, uses parquet format, supports multiple data libraries, and is described as multimoda... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset metadata into a concise SFT training example. | 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 through the Vicuna demo and Chatbot Arena website fro... | LMSYS-Chat-1M is a real-world conversation dataset on Hugging Face with one million chats involving 25 modern LLMs. It was collected from Vicuna demo and Chatbot Arena traffic across 210,000 unique IP addresses between April and August 2023. Each record includes structured conversation data, model metadata, language ta... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise description for supervised fine-tuning data curation. | Dataset 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... | ShareGPT_Vicuna_unfiltered is an English-language Hugging Face dataset intended for conversational model training and cleaning workflows. It is a refined version of ShareGPT/Vicuna-style data, with options for variants that either remove or retain instances of the phrase "I'm sorry, but". The dataset is licensed under ... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise description for supervised fine-tuning data curation. | Dataset 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,... | finepdfs is a massive multilingual Hugging Face dataset built exclusively from PDF sources. It contains approximately 3 trillion tokens drawn from 475 million documents spanning 1,733 languages, making it one of the largest publicly available PDF-derived corpora for text generation and large-scale language model traini... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset based on the provided metadata. | Name: PhysicalAI-Autonomous-Vehicles
Source: huggingface
ID: nvidia/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 gen... | PhysicalAI-Autonomous-Vehicles is a Hugging Face dataset from NVIDIA for autonomous vehicle research. It contains a large, geographically diverse multi-sensor driving collection designed to support end-to-end Physical AI driving systems. The dataset includes about 1700 hours of driving data, with both collection and la... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset based on the provided metadata. | Name: OpenThoughts-114k
Source: huggingface
ID: open-thoughts/OpenThoughts-114k
Description: Open-Thoughts-114k. Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. Available subsets include a default subset containing ready-to-train data used to finetune the Open... | OpenThoughts-114k is a Hugging Face synthetic reasoning dataset with 114000 high-quality text examples spanning math, science, code, and puzzles. It is distributed in Parquet format, supports multiple data libraries including datasets, Dask, Polars, and mlcroissant, and includes a ready-to-train default subset used to ... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction tuning, including its purpose, scale, format, language, notable characteristics, and popularity signals. | Name: OpenHermes-2.5
Source: huggingface
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 underpinning the Open Hermes 2/2.5 an... | OpenHermes-2.5 is a large English text dataset on Hugging Face intended for instruction-tuning and general LLM training. It is distributed in JSON format and falls in the 1M to 10M sample size range. The dataset is presented as the core compilation used to build OpenHermes 2.5 and Nous Hermes 2 model series, combining ... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction tuning, including its purpose, scale, format, language, cleaning goals, and popularity signals. | Name: alpaca-cleaned
Source: huggingface
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 hallucination-prone instructions that referenced internet data.
Tags: task_categories:text-generatio... | alpaca-cleaned is an English text-generation dataset on Hugging Face designed for instruction fine-tuning. It is a cleaned version of Stanford's original Alpaca dataset, with fixes aimed at removing problematic examples, especially instructions that encouraged hallucinations by referring to internet-based information u... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset metadata into a concise SFT training example. | Name: fineweb-2
Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining data for over 1000 languages. It is reproducible, released under the ODC-By 1.0 license, and validated through extensive ablation experiments.
... | {"name":"fineweb-2","platform":"Hugging Face","dataset_id":"HuggingFaceFW/fineweb-2","summary":"FineWeb2 is a large-scale multilingual pretraining dataset covering over 1000 languages, designed for high-quality and reproducible text generation research.","task_category":"text-generation","modalities":["tabular","text"]... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset metadata into a concise SFT training example. | Name: hle
Source: huggingface
Dataset ID: cais/hle
Description: Humanity's Last Exam (HLE) is a multi-modal benchmark at the frontier of human knowledge, designed as a closed-ended academic benchmark with broad subject coverage. It contains 2,500 questions and includes image and text modalities.
Tags: benchmark:officia... | {"name":"hle","platform":"Hugging Face","dataset_id":"cais/hle","summary":"Humanity's Last Exam is an official multimodal benchmark with 2,500 closed-ended questions spanning broad academic subjects at the frontier of human knowledge.","type":"benchmark","modalities":["image","text"],"format":"parquet","license":"mit",... | huggingface | cais/hle |
Summarize the dataset and identify its main machine learning use case. | 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 synset. There are more than ... | ImageNet-1k is a large-scale image dataset from the ILSVRC 2012 benchmark, structured around WordNet synsets. It is primarily used for multi-class image classification, where models learn to classify images into one of many object categories. The dataset is crowdsourced, English-labeled, image-based, and widely used fo... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset and describe what kind of model training it supports. | Name: Alpaca-CoT
Source: huggingface
Dataset ID: QingyiSi/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. The project also includes empirical study on various instruction-tun... | Alpaca-CoT is a collection of standardized instruction-tuning datasets designed for supervised fine-tuning of instruction-following language models. It supports training models like Alpaca on multilingual instruction data, including chain-of-thought style examples. The dataset is useful for improving response quality, ... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset and highlight its purpose, scale, languages, tasks, format, license, and notable characteristics. | Dataset 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 lan... | PersonaHub is a Hugging Face dataset focused on large-scale synthetic data creation through persona-driven generation. It accompanies the paper "Scaling Synthetic Data Creation with 1,000,000,000 Personas" and introduces PERSONA HUB, a collection of 1 billion diverse personas automatically curated from web data to supp... | huggingface | proj-persona/PersonaHub |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"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 small open prompt collection mirrored from the prompts.chat community platform. It is best suited for prompt engineering research, prompt retrieval, instruction template discovery, chatbot ideation, and benchmarking prompt diversity across popular LLM ecosystems such as ChatGPT, Claude, Gemini, Llama,... | huggingface | fka/prompts.chat |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"HuggingFaceFW/fineweb","name":"fineweb","description":"FineWeb: 15 trillion tokens of the finest data the web has to offer. The FineWeb dataset consists of more than 18.5T tokens of cleaned and deduplicated English web data from CommonCrawl. The data processing pipeline is optimized for LL... | FineWeb is a massive English web-scale corpus built from cleaned and deduplicated CommonCrawl data, intended primarily for large language model pretraining. Its defining characteristic is scale: over 18.5 trillion tokens, making it far larger than typical supervised fine-tuning datasets. It is useful for continued pret... | huggingface | HuggingFaceFW/fineweb |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"Anthropic/hh-rlhf","name":"hh-rlhf","description":"Dataset Card for HH-RLHF. 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 t... | HH-RLHF is a human feedback dataset focused on helpfulness and harmlessness preferences, designed for training reward models or preference models in RLHF pipelines. It is a medium-to-large text dataset in JSON format with roughly 100K to 1M examples and an MIT license. Its strongest use cases are preference learning, r... | huggingface | Anthropic/hh-rlhf |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"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 outlined in the Orca paper. It has been instrumental in generating high-performing model checkpoints and serves as a v... | OpenOrca is a large English instruction-style dataset built from augmented FLAN-style data and inspired by the Orca paper’s distribution. It is highly relevant for supervised fine-tuning because it covers a broad range of NLP tasks, including question answering, summarization, classification, feature extraction, and te... | huggingface | Open-Orca/OpenOrca |
Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability. | {"source":"huggingface","id":"OpenAssistant/oasst1","name":"oasst1","description":"OpenAssistant Conversations Dataset (OASST1). A human-generated, human-annotated assistant-style conversation corpus consisting of 161,443 messages in 35 different languages, annotated with 461,292 quality ratings, resulting in over 10,0... | OASST1 is a multilingual, human-generated and human-annotated assistant conversation dataset created for alignment research. It contains over 161K messages across 35 languages, along with quality ratings and conversation tree structure, making it especially valuable for supervised fine-tuning of chat assistants, multil... | huggingface | OpenAssistant/oasst1 |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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. The dataset was created to support the task of question answering on basic mathematical problems that r... | gsm8k is a Hugging Face dataset focused on grade school math word problems that require multi-step reasoning. It is intended for question answering and text-generation style evaluation on basic arithmetic reasoning. The data is text-only, monolingual English, and tagged as an official benchmark with original crowdsourc... | huggingface | openai/gsm8k |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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 the Wikipedia dumps with one subset per language, each containing a single train split. Each example contai... | wikipedia is a Hugging Face dataset containing cleaned full Wikipedia articles across many languages, derived from Wikimedia dumps. It is primarily useful for language modeling, masked language modeling, text generation, and fill-mask tasks. The modality is text, and the dataset is massively multilingual with one subse... | huggingface | wikimedia/wikipedia |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"source":"huggingface","id":"gsdf/EasyNegative","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"... | EasyNegative is a Hugging Face image-related asset described as a negative embedding trained with Counterfeit for use with Stable Diffusion WebUI embeddings. Its modality is image, and it appears intended for generative image workflows rather than conventional NLP tasks. Notable metadata includes imagefolder format, ve... | huggingface | gsdf/EasyNegative |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"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 Hugging Face text dataset positioned as a clean-room, fully open-source implementation of the LLaMa dataset. Its main purpose is text generation and large-scale language model training, and the metadata indicates English text. Notable tags show text modality, large size, and compatibility with da... | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the following dataset metadata for supervised fine-tuning. Include the dataset name, source, main purpose, modality, languages if clear, notable tags, popularity signals, and any important caveats. | {"source":"huggingface","id":"FreedomIntelligence/medical-o1-reasoning-SFT","name":"medical-o1-reasoning-SFT","description":"News: [2025/04/22] We split the data and kept only the medical SFT dataset (medical_o1_sft.json). The file medical_o1_sft_mix.json contains a mix of medical and general instruction data. [2025/02... | medical-o1-reasoning-SFT is a Hugging Face dataset for medical reasoning supervised fine-tuning, built around medical verifiable problems and distilled reasoning data. It supports question answering and text generation tasks in text modality, with English and Chinese language coverage. Notable metadata includes Apache-... | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"HuggingFaceFW/fineweb-edu","name":"fineweb-edu","description":"1.3 trillion tokens of the finest educational data the web has to offer. FineWeb-Edu consists of 1.3T tokens of educational web pages filtered from FineWeb using an educational quality classifier with annotations generated by L... | fineweb-edu is a Hugging Face dataset for text generation focused on high-quality educational web content in English. It contains about 1.3 trillion tokens filtered from FineWeb using an educational quality classifier, making it useful for large-scale language model pretraining. The dataset is available in parquet form... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"allenai/dolma","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-m... | dolma is a Hugging Face dataset for text generation and language model pretraining research. It is an open English corpus containing roughly three trillion tokens, designed for large-scale LLM training and language modeling use cases. The dataset is licensed under ODC-By and references papers such as arXiv:2402.00159. ... | huggingface | allenai/dolma |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"bigcode/the-stack","name":"the-stack","description":"Dataset Card for The Stack. Initial release included 30 programming languages and 18 permissive licenses, with a near-deduplicated dataset size of 3TB. Later versions excluded weak copyleft licenses and expanded the permissive license li... | the-stack is a Hugging Face dataset for text generation over source code. It is a multilingual code corpus built from crowdsourced and expert-generated repositories, near-deduplicated, and originally around 3TB in size. The dataset is distributed in parquet format with text and tabular modalities, uses a nonstandard or... | huggingface | bigcode/the-stack |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"roneneldan/TinyStories","name":"TinyStories","description":"Dataset containing synthetically generated by GPT-3.5 and GPT-4 short stories that only use a small vocabulary. Described in arXiv:2305.07759 and used to train small language models.","tags":["task_categories:text-generation","lan... | TinyStories is a Hugging Face dataset for English text generation composed of synthetic short stories generated by GPT-3.5 and GPT-4 using a restricted vocabulary. It is intended for training and studying small language models, with a dataset size in the 1M to 10M range. The dataset is available in parquet format, cont... | huggingface | roneneldan/TinyStories |
Summarize the dataset metadata into a concise SFT-style description including source, name, purpose, scale, modalities, license, notable tags, popularity, and any referenced paper. | {"source":"huggingface","id":"databricks/databricks-dolly-15k","name":"databricks-dolly-15k","description":"databricks-dolly-15k is an open source dataset of instruction-following records generated by thousands of Databricks employees across categories including brainstorming, classification, closed QA, generation, inf... | databricks-dolly-15k is a Hugging Face instruction-tuning dataset in English containing open-source instruction-following examples created by Databricks employees. It covers tasks such as question answering, summarization, brainstorming, classification, generation, and information extraction, with a size between 10K an... | huggingface | databricks/databricks-dolly-15k |
Summarize the given dataset entry for instruction-tuning relevance, modality, language, license, scale, and notable metadata. | Name: alpaca
Source: huggingface
ID: tatsu-lab/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. The a... | Alpaca is an English text dataset on Hugging Face designed for instruction fine-tuning. It contains about 52,000 instruction-response demonstrations generated with text-davinci-003 and based on the Self-Instruct pipeline. The dataset is tagged for text generation, uses Parquet format, and falls in the 10K to 100K size ... | huggingface | tatsu-lab/alpaca |
Summarize the given dataset entry, highlighting what it is for, modality, usage context, license, size, and key tags. | Name: bad_prompt
Source: huggingface
ID: Nerfgun3/bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to train the negative prompt as an embedding, unifying the basis of the negative prompt into one word or embedding. It is noted as helpful for image generation, especially... | bad_prompt is a small English Hugging Face image-related resource associated with Stable Diffusion workflows. It is described as a negative embedding or textual inversion asset intended to compress negative prompt behavior into a reusable embedding, with claimed benefits for generation quality such as improved hands. T... | huggingface | Nerfgun3/bad_prompt |
Summarize the dataset metadata into a concise SFT training example. | 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 built through stringent filtering and large-scale deduplication of CommonCrawl. The dataset is described as achiev... | Falcon RefinedWeb is a large English web text dataset from TII on Hugging Face. It is created from CommonCrawl using strict filtering and large-scale deduplication, and is licensed under ODC-By 1.0. The dataset targets text generation, uses parquet format, supports multiple data libraries, and is described as multimoda... | huggingface | tiiuae/falcon-refinedweb |
Summarize the dataset metadata into a concise SFT training example. | 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 through the Vicuna demo and Chatbot Arena website fro... | LMSYS-Chat-1M is a real-world conversation dataset on Hugging Face with one million chats involving 25 modern LLMs. It was collected from Vicuna demo and Chatbot Arena traffic across 210,000 unique IP addresses between April and August 2023. Each record includes structured conversation data, model metadata, language ta... | huggingface | lmsys/lmsys-chat-1m |
Summarize the dataset metadata into a concise description for supervised fine-tuning data curation. | Dataset 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... | ShareGPT_Vicuna_unfiltered is an English-language Hugging Face dataset intended for conversational model training and cleaning workflows. It is a refined version of ShareGPT/Vicuna-style data, with options for variants that either remove or retain instances of the phrase "I'm sorry, but". The dataset is licensed under ... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset metadata into a concise description for supervised fine-tuning data curation. | Dataset 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,... | finepdfs is a massive multilingual Hugging Face dataset built exclusively from PDF sources. It contains approximately 3 trillion tokens drawn from 475 million documents spanning 1,733 languages, making it one of the largest publicly available PDF-derived corpora for text generation and large-scale language model traini... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset based on the provided metadata. | Name: PhysicalAI-Autonomous-Vehicles
Source: huggingface
ID: nvidia/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 gen... | PhysicalAI-Autonomous-Vehicles is a Hugging Face dataset from NVIDIA for autonomous vehicle research. It contains a large, geographically diverse multi-sensor driving collection designed to support end-to-end Physical AI driving systems. The dataset includes about 1700 hours of driving data, with both collection and la... | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset based on the provided metadata. | Name: OpenThoughts-114k
Source: huggingface
ID: open-thoughts/OpenThoughts-114k
Description: Open-Thoughts-114k. Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. Available subsets include a default subset containing ready-to-train data used to finetune the Open... | OpenThoughts-114k is a Hugging Face synthetic reasoning dataset with 114000 high-quality text examples spanning math, science, code, and puzzles. It is distributed in Parquet format, supports multiple data libraries including datasets, Dask, Polars, and mlcroissant, and includes a ready-to-train default subset used to ... | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for instruction tuning, including its purpose, scale, format, language, notable characteristics, and popularity signals. | Name: OpenHermes-2.5
Source: huggingface
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 underpinning the Open Hermes 2/2.5 an... | OpenHermes-2.5 is a large English text dataset on Hugging Face intended for instruction-tuning and general LLM training. It is distributed in JSON format and falls in the 1M to 10M sample size range. The dataset is presented as the core compilation used to build OpenHermes 2.5 and Nous Hermes 2 model series, combining ... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for instruction tuning, including its purpose, scale, format, language, cleaning goals, and popularity signals. | Name: alpaca-cleaned
Source: huggingface
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 hallucination-prone instructions that referenced internet data.
Tags: task_categories:text-generatio... | alpaca-cleaned is an English text-generation dataset on Hugging Face designed for instruction fine-tuning. It is a cleaned version of Stanford's original Alpaca dataset, with fixes aimed at removing problematic examples, especially instructions that encouraged hallucinations by referring to internet-based information u... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset metadata into a concise SFT training example. | Name: fineweb-2
Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality pretraining data for over 1000 languages. It is reproducible, released under the ODC-By 1.0 license, and validated through extensive ablation experiments.
... | {"name":"fineweb-2","platform":"Hugging Face","dataset_id":"HuggingFaceFW/fineweb-2","summary":"FineWeb2 is a large-scale multilingual pretraining dataset covering over 1000 languages, designed for high-quality and reproducible text generation research.","task_category":"text-generation","modalities":["tabular","text"]... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset metadata into a concise SFT training example. | Name: hle
Source: huggingface
Dataset ID: cais/hle
Description: Humanity's Last Exam (HLE) is a multi-modal benchmark at the frontier of human knowledge, designed as a closed-ended academic benchmark with broad subject coverage. It contains 2,500 questions and includes image and text modalities.
Tags: benchmark:officia... | {"name":"hle","platform":"Hugging Face","dataset_id":"cais/hle","summary":"Humanity's Last Exam is an official multimodal benchmark with 2,500 closed-ended questions spanning broad academic subjects at the frontier of human knowledge.","type":"benchmark","modalities":["image","text"],"format":"parquet","license":"mit",... | huggingface | cais/hle |
Summarize the dataset and identify its main machine learning use case. | 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 synset. There are more than ... | ImageNet-1k is a large-scale image dataset from the ILSVRC 2012 benchmark, structured around WordNet synsets. It is primarily used for multi-class image classification, where models learn to classify images into one of many object categories. The dataset is crowdsourced, English-labeled, image-based, and widely used fo... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset and describe what kind of model training it supports. | Name: Alpaca-CoT
Source: huggingface
Dataset ID: QingyiSi/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. The project also includes empirical study on various instruction-tun... | Alpaca-CoT is a collection of standardized instruction-tuning datasets designed for supervised fine-tuning of instruction-following language models. It supports training models like Alpaca on multilingual instruction data, including chain-of-thought style examples. The dataset is useful for improving response quality, ... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset and highlight its purpose, scale, languages, tasks, format, license, and notable characteristics. | Dataset 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 lan... | PersonaHub is a Hugging Face dataset focused on large-scale synthetic data creation through persona-driven generation. It accompanies the paper "Scaling Synthetic Data Creation with 1,000,000,000 Personas" and introduces PERSONA HUB, a collection of 1 billion diverse personas automatically curated from web data to supp... | huggingface | proj-persona/PersonaHub |
SoloAI SFT Dataset: 20260424-2139
📊 数据集概览
| 属性 | 值 |
|---|---|
| 数据条数 | 30 条 |
| 数据源 | huggingface |
| 生成时间 | 2026-04-24 21:39 |
| 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': 'Given a raw dataset catalog entry, summarize it for LLM supervised fine-tuning selection. Include what the dataset is, likely use cases, notable scale or language coverage, and any important caution about suitability.', '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","region:us","ChatGPT","prompts","AI","GPT","Claude","Gemini","Llama","Mistral","LLM","prompt-engineering","conversational-ai","text-generation","chatbot","awesome-list"],"downloads":44569,"likes":9680,"created_at":""}', 'output': 'prompts.chat is a small open prompt collection mirrored from the prompts.chat community platform. It is best suited for prompt engineering research, prompt retrieval, instruction template discovery, chatbot ideation, and benchmarking prompt diversity across popular LLM ecosystems such as ChatGPT, Claude, Gemini, Llama, and Mistral. The dataset is text-based, distributed as CSV, licensed under CC0-1.0, and falls in the 1K to 10K size range. A key caution is that this appears to be a prompt repository rather than a curated high-quality instruction-response SFT corpus, so it is more useful for prompt libraries, seed instruction generation, or prompt analysis than for direct assistant fine-tuning unless additional responses or filtering are added.', '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 | 2026-04-24 21:39 | 初始发布，30 条数据 |
- Downloads last month
- 14