instruction stringlengths 53 167 | input stringlengths 212 1.04k | output stringlengths 214 675 | source stringclasses 1
value | original_id stringlengths 8 45 |
|---|---|---|---|---|
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: fka/prompts.chat
name: prompts.chat
description: a.k.a. Awesome ChatGPT Prompts. This is a mirror of prompts.chat, a social platform for AI prompts. It is an open-source platform where users can share, discover, and collect AI prompts from the community.
tags: task_categories:q... | prompts.chat is a community prompt collection mirrored from the prompts.chat platform, also known as Awesome ChatGPT Prompts. It is primarily useful for prompt engineering research, prompt library creation, chatbot experimentation, and studying how users structure prompts for systems like ChatGPT, Claude, Gemini, Llama... | huggingface | fka/prompts.chat |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: HuggingFaceFW/fineweb
name: fineweb
description: FineWeb is a large cleaned and deduplicated English web dataset from CommonCrawl, consisting of more than 18.5T tokens. The processing pipeline is optimized for LLM performance and built with datatrove.
tags: task_categories:text... | FineWeb is a massive English web-text dataset derived from CommonCrawl and optimized for large language model training through cleaning and deduplication. It is especially suited for large-scale pretraining, continued pretraining, corpus analysis, and data quality research for text generation systems. The dataset is ex... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: Anthropic/hh-rlhf
name: hh-rlhf
description: This repository provides human preference data about helpfulness and harmlessness for training preference or reward models in RLHF. The card explicitly notes the data are not meant for supervised training of dialogue agents.
tags: li... | HH-RLHF is a human feedback dataset focused on helpfulness and harmlessness preferences, designed primarily for training reward models and supporting RLHF pipelines. It is not intended as a standard supervised fine-tuning dialogue dataset, and the dataset card explicitly warns against using it directly for that purpose... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: Open-Orca/OpenOrca
name: OpenOrca
description: OpenOrca 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 high-performing model checkpoints.
tags: task_categories:text-clas... | OpenOrca is a large instruction-style dataset built from augmented FLAN data and designed to reflect distributions from the Orca line of work. It is well suited for supervised fine-tuning of general-purpose assistants, especially across tasks such as question answering, summarization, classification, zero-shot tasks, f... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
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 tre... | OASST1 is a multilingual assistant conversation dataset built from human-generated and human-annotated dialogue trees. It is highly useful for supervised fine-tuning of chat assistants, multilingual dialogue modeling, response ranking, and alignment research thanks to both conversation structure and quality ratings. Th... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset and identify its main use case. | Name: gsm8k
Description: 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 requiring multi-step reasoning. Problems typically take between 2 and 8 steps to solve and mainly involve el... | GSM8K is an English math word-problem dataset containing about 8.5K grade-school-level questions. It is designed for evaluating and training models on multi-step mathematical reasoning and question answering, especially problems that require several arithmetic operations to reach the final answer. | huggingface | openai/gsm8k |
Summarize the dataset and identify its main use case. | Name: wikipedia
Description: Wikipedia dataset containing cleaned articles of all languages. It is built from Wikimedia dumps with one subset per language, each containing a single train split. Each example contains the content of one full Wikipedia article with cleaning to strip markdown and unwanted sections.
Tags: t... | The Wikimedia Wikipedia dataset is a large multilingual text corpus made of cleaned full Wikipedia articles across many languages. Its main use case is language model pretraining and related tasks such as text generation and masked language modeling. | huggingface | wikimedia/wikipedia |
Summarize the dataset and identify its main use case. | Name: EasyNegative
Description: This is a negative embedding trained with Counterfeit. It is intended for use in the stable-diffusion-webui embeddings folder and may also work with other models, though effectiveness is uncertain.
Tags: license:other, format:imagefolder, modality:image
Downloads: 31720
Likes: 1174 | EasyNegative is an image-related resource described as a negative embedding trained with Counterfeit for Stable Diffusion workflows. Its main use case is improving or steering image generation by serving as a negative embedding in diffusion-based image generation pipelines. | huggingface | gsdf/EasyNegative |
Summarize the dataset and identify its main use case. | 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, modality:text
Downloads: 2169
Likes: 1147 | RedPajama-Data-1T is a large English text dataset intended as an open-source reproduction of the data used for LLaMA-style models. Its primary use case is large-scale language model pretraining for text generation. | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset and identify its main use case. | Name: medical-o1-reasoning-SFT
Description: A medical reasoning dataset for supervised fine-tuning built on medical verifiable problems and distilled reasoning data. It includes medical SFT data and related releases for reasoning-oriented model training.
Tags: task_categories:question-answering, task_categories:text-ge... | medical-o1-reasoning-SFT is a bilingual English-Chinese medical reasoning dataset designed for supervised fine-tuning. Its main use case is training models to answer medical questions and generate step-by-step reasoning on medically verifiable problems. | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-edu
Name: fineweb-edu
Description: 📚 FineWeb-Edu. 1.3 trillion tokens of the finest educational data the web has to offer. FineWeb-Edu consists of 1.3T tokens and 5.4T tokens (FineWeb-Edu-score-2) of educational web pages filtered from FineWeb dataset. This is the ... | fineweb-edu is a large-scale English educational text dataset on Hugging Face. It contains about 1.3 trillion tokens of web data filtered from FineWeb for educational quality, using a classifier trained from Llama3-70B-Instruct-generated annotations. The dataset is intended for text generation and language model pretra... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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-modeling, casual... | dolma is an open English text corpus designed for language model pretraining research. It contains roughly three trillion tokens and is positioned as a large-scale resource for text generation and LLM development. The dataset is hosted on Hugging Face and released under the ODC-By license. | huggingface | allenai/dolma |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset ID: bigcode/the-stack
Name: the-stack
Description: Dataset Card for The Stack. Initial release included 30 programming languages and 18 permissive licenses, producing a near-deduplicated dataset of 3TB. Later versions excluded weak copyleft licenses (MPL/EPL/LGPL) and expanded the list of pe... | the-stack is a multilingual code dataset for text generation and code model training. It contains source code across many programming languages, is near-deduplicated, and was curated to emphasize permissive licensing. The dataset is available on Hugging Face in parquet format and supports large-scale code LLM research. | huggingface | bigcode/the-stack |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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 paper: https://arxiv.org/abs/2305.07759. The models referred to in the paper were trained on TinyStories-tr... | TinyStories is an English text dataset of synthetically generated short stories created by GPT-3.5 and GPT-4 using a limited vocabulary. It is designed for text generation research, especially for training small language models on simple narrative text. The dataset is hosted on Hugging Face and released under the CDLA-... | huggingface | roneneldan/TinyStories |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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 in behavioral categories outlined in the InstructGPT paper, including brainstorming, clas... | databricks-dolly-15k is an English instruction-following dataset containing about 15,000 examples created by Databricks employees. It covers tasks such as brainstorming, classification, question answering, information extraction, generation, and summarization. The dataset is well suited for supervised fine-tuning of in... | huggingface | databricks/databricks-dolly-15k |
Summarize the dataset metadata and explain why this dataset is useful for supervised fine-tuning. | 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-generation dataset containing about 52,000 instruction-response examples. It was generated using OpenAI's text-davinci-003 and designed specifically for instruction tuning, making it highly relevant for supervised fine-tuning of language models. The dataset is in parquet format, text modality,... | huggingface | tatsu-lab/alpaca |
Analyze the dataset card and describe the dataset's purpose, modality, and likely training use case. | Name: bad_prompt
Source: huggingface
ID: Nerfgun3/bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to somehow train the negative prompt as an embedding, thus unifying the basis of the negative prompt into one word or embedding. Embedding has proven to be very helpful fo... | bad_prompt is a small image-related dataset or asset associated with Stable Diffusion workflows, especially negative embedding or textual inversion usage. Its modality is image, and it appears intended to improve image generation quality by encoding a reusable negative prompt concept into an embedding. The likely use c... | huggingface | Nerfgun3/bad_prompt |
Provide a concise technical summary of this dataset, including scale, content type, and potential machine learning applications. | Name: falcon-refinedweb
Source: huggingface
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. Models trained on RefinedWeb achieve perfo... | Falcon RefinedWeb is a very large English web-text dataset derived from CommonCrawl with aggressive filtering and deduplication. It is text modality, stored in parquet format, and its size category indicates hundreds of millions to under one billion records. The dataset is primarily useful for large-scale pretraining o... | huggingface | tiiuae/falcon-refinedweb |
Read the metadata and explain what this dataset contains and how it could be used in LLM training or evaluation. | Name: lmsys-chat-1m
Source: huggingface
ID: lmsys/lmsys-chat-1m
Description: This dataset contains one million real-world conversations with 25 state-of-the-art LLMs. It was collected from 210K unique IP addresses on the Vicuna demo and Chatbot Arena website from April to August 2023. Each sample includes a conversatio... | LMSYS-Chat-1M is a large real-world conversation dataset containing one million chats involving 25 LLMs. It captures authentic user-model interactions collected from public demo platforms and includes structured conversation text, metadata such as model name, language, and moderation labels. This makes it valuable for ... | huggingface | lmsys/lmsys-chat-1m |
Summarize this dataset and describe its likely role in instruction-tuning or chatbot model development. | Name: ShareGPT_Vicuna_unfiltered
Source: huggingface
ID: anon8231489123/ShareGPT_Vicuna_unfiltered
Description: Further cleaning done. Please look through the dataset and ensure that I didn't miss anything. Confirmed working method for training the model. Two choices are provided: one version removes instances of "I'm ... | ShareGPT_Vicuna_unfiltered appears to be an English conversational dataset derived from ShareGPT-style chat data and prepared for Vicuna-like model training. The description suggests it has undergone some cleaning and offers variants that either remove or keep common refusal phrases such as "I'm sorry, but". Its likely... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset for supervised fine-tuning use. | Name: finepdfs
Description: Liberating 3T of the finest tokens from PDFs. FinePDFs is a publicly available corpus sourced exclusively from PDFs, containing about 3 trillion tokens across 475 million documents in 1733 languages.
Tags: task_categories:text-generation, license:odc-by, size_categories:100M<n<1B, format:par... | finepdfs is a massive multilingual text dataset from PDFs on Hugging Face. It contains roughly 3 trillion tokens from 475 million documents spanning 1733 languages. The dataset is geared toward text generation and distributed in Parquet format, with support for tools like datasets, Dask, and Polars. Its license is ODC-... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset for supervised fine-tuning use. | Name: PhysicalAI-Autonomous-Vehicles
Description: The PhysicalAI-Autonomous-Vehicles dataset provides one of the largest, geographically diverse collections of multi-sensor data for AV researchers to build end-to-end driving systems. It is ready for commercial/non-commercial AV use per the license agreement. The datase... | PhysicalAI-Autonomous-Vehicles is a large-scale autonomous driving dataset from NVIDIA. It offers geographically diverse multi-sensor driving data and includes about 1700 hours of driving. The dataset is intended for building Physical AI and end-to-end driving systems, with usage governed by a custom license agreement. | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset for supervised fine-tuning use. | Name: OpenThoughts-114k
Description: Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. The default subset contains ready-to-train data used to finetune OpenThinker models.
Tags: license:apache-2.0, size_categories:100K<n<1M, format:parquet, modality:text, librar... | OpenThoughts-114k is a synthetic reasoning dataset with 114,000 high-quality training examples. It covers domains such as mathematics, science, coding, and puzzles, and is designed for direct model fine-tuning. The dataset is text-based, distributed in Parquet format, and released under the Apache 2.0 license. | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for supervised fine-tuning use. | Name: alpaca-cleaned
Description: This is a cleaned version of the original Alpaca dataset released by Stanford. It fixes issues such as hallucination-prone instructions and other quality problems in the original release.
Tags: task_categories:text-generation, language:en, license:cc-by-4.0, size_categories:10K<n<100K,... | alpaca-cleaned is an English instruction-tuning dataset derived from the original Stanford Alpaca release. It removes or fixes problematic examples, especially prompts that could trigger hallucinated answers due to missing internet context. The dataset is distributed in JSON format and is commonly used for instruction ... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset for supervised fine-tuning use. | Name: OpenHermes-2.5
Description: This is the dataset that made OpenHermes 2.5 and Nous Hermes 2 series of models. It is a compilation and curation of many open source datasets and custom created synthetic datasets.
Tags: language:eng, size_categories:1M<n<10M, format:json, modality:text, library:datasets, library:pand... | OpenHermes-2.5 is a large English text dataset used to train the OpenHermes 2.5 and Nous Hermes 2 model families. It is a curated compilation of open-source and synthetic data, including GPT-4 distilled examples. The dataset is delivered in JSON format and is aimed at general instruction and assistant-style model train... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for supervised fine-tuning use. Include its purpose, modality, scale, languages, license, and notable characteristics. | Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality reproducible pretraining data across more than 1000 languages. It is extensively validated and available under the ODC-By 1.0 license.
Tags: task_categories:text-generation; many language tags; license:odc-by; s... | fineweb-2 is a large-scale multilingual text dataset intended for text generation and language model pretraining. It is the second version of FineWeb and emphasizes high-quality, fully reproducible data collection and processing. The dataset covers over 1000 languages, making it especially notable for broad multilingua... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset for supervised fine-tuning use. Include its purpose, modality, scale, license, and any special usage considerations. | Name: hle
Description: Humanity's Last Exam is a multimodal benchmark designed as a frontier closed-ended academic benchmark spanning broad subjects, with 2,500 questions. The dataset page warns users not to publicly share, re-upload, or distribute the dataset.
Tags: benchmark:official; license:mit; size_categories:1K<... | hle, short for Humanity's Last Exam, is an official multimodal benchmark designed to evaluate models on frontier-level academic knowledge through closed-ended questions. It contains about 2,500 questions and spans many subject areas. The dataset includes both image and text modalities, is distributed in parquet format,... | huggingface | cais/hle |
Summarize the dataset for supervised fine-tuning use. Include its task, modality, scale, language, licensing situation, and notable background. | Name: imagenet-1k
Description: ImageNet, specifically ILSVRC 2012, is an image dataset organized according to the WordNet hierarchy. It contains images illustrating synsets and is widely used for image classification.
Tags: task_categories:image-classification; task_ids:multi-class-image-classification; language:en; li... | imagenet-1k is a large-scale image classification dataset based on the ILSVRC 2012 subset of ImageNet. It organizes visual concepts according to the WordNet hierarchy and supports multi-class image classification. The dataset is monolingual in English for labels and metadata, uses image modality, and is in the 1M to 10... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset for supervised fine-tuning use. Include its purpose, language coverage, license, and why it is useful for instruction tuning. | Name: Alpaca-CoT
Description: A continuously collected instruction-finetuning dataset collection that standardizes multiple instruction-tuning datasets into a common format directly loadable by Alpaca-style code. It is associated with empirical studies on instruction-tuning datasets.
Tags: language:en; language:zh; lan... | Alpaca-CoT is an instruction-finetuning dataset collection designed to aggregate and standardize multiple instruction-tuning sources into a unified format compatible with Alpaca-style training pipelines. It is especially useful for supervised fine-tuning because it focuses on instruction-following data and chain-of-tho... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset for supervised fine-tuning use. Include its purpose, language, task types, size, license, and composition highlights. | Name: Chinese-DeepSeek-R1-Distill-data-110k
Description: A Chinese open-source distilled dataset based on DeepSeek-R1, intended for SFT use. It provides a version where reasoning and answers are merged into an output field for direct loading by many SFT frameworks. It contains 110K samples spanning math and general-pur... | Chinese-DeepSeek-R1-Distill-data-110k is a Chinese supervised fine-tuning dataset distilled from DeepSeek-R1-style outputs. It is designed specifically for SFT and even provides a directly usable version in which reasoning and final answers are merged into a single output field. The dataset supports text generation and... | huggingface | Congliu/Chinese-DeepSeek-R1-Distill-data-110k |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: fka/prompts.chat
name: prompts.chat
description: a.k.a. Awesome ChatGPT Prompts. This is a mirror of prompts.chat, a social platform for AI prompts. It is an open-source platform where users can share, discover, and collect AI prompts from the community.
tags: task_categories:q... | prompts.chat is a community prompt collection mirrored from the prompts.chat platform, also known as Awesome ChatGPT Prompts. It is primarily useful for prompt engineering research, prompt library creation, chatbot experimentation, and studying how users structure prompts for systems like ChatGPT, Claude, Gemini, Llama... | huggingface | fka/prompts.chat |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: HuggingFaceFW/fineweb
name: fineweb
description: FineWeb is a large cleaned and deduplicated English web dataset from CommonCrawl, consisting of more than 18.5T tokens. The processing pipeline is optimized for LLM performance and built with datatrove.
tags: task_categories:text... | FineWeb is a massive English web-text dataset derived from CommonCrawl and optimized for large language model training through cleaning and deduplication. It is especially suited for large-scale pretraining, continued pretraining, corpus analysis, and data quality research for text generation systems. The dataset is ex... | huggingface | HuggingFaceFW/fineweb |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: Anthropic/hh-rlhf
name: hh-rlhf
description: This repository provides human preference data about helpfulness and harmlessness for training preference or reward models in RLHF. The card explicitly notes the data are not meant for supervised training of dialogue agents.
tags: li... | HH-RLHF is a human feedback dataset focused on helpfulness and harmlessness preferences, designed primarily for training reward models and supporting RLHF pipelines. It is not intended as a standard supervised fine-tuning dialogue dataset, and the dataset card explicitly warns against using it directly for that purpose... | huggingface | Anthropic/hh-rlhf |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
source: huggingface
id: Open-Orca/OpenOrca
name: OpenOrca
description: OpenOrca 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 high-performing model checkpoints.
tags: task_categories:text-clas... | OpenOrca is a large instruction-style dataset built from augmented FLAN data and designed to reflect distributions from the Orca line of work. It is well suited for supervised fine-tuning of general-purpose assistants, especially across tasks such as question answering, summarization, classification, zero-shot tasks, f... | huggingface | Open-Orca/OpenOrca |
Summarize the dataset for instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags. | Dataset metadata:
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 tre... | OASST1 is a multilingual assistant conversation dataset built from human-generated and human-annotated dialogue trees. It is highly useful for supervised fine-tuning of chat assistants, multilingual dialogue modeling, response ranking, and alignment research thanks to both conversation structure and quality ratings. Th... | huggingface | OpenAssistant/oasst1 |
Summarize the dataset and identify its main use case. | Name: gsm8k
Description: 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 requiring multi-step reasoning. Problems typically take between 2 and 8 steps to solve and mainly involve el... | GSM8K is an English math word-problem dataset containing about 8.5K grade-school-level questions. It is designed for evaluating and training models on multi-step mathematical reasoning and question answering, especially problems that require several arithmetic operations to reach the final answer. | huggingface | openai/gsm8k |
Summarize the dataset and identify its main use case. | Name: wikipedia
Description: Wikipedia dataset containing cleaned articles of all languages. It is built from Wikimedia dumps with one subset per language, each containing a single train split. Each example contains the content of one full Wikipedia article with cleaning to strip markdown and unwanted sections.
Tags: t... | The Wikimedia Wikipedia dataset is a large multilingual text corpus made of cleaned full Wikipedia articles across many languages. Its main use case is language model pretraining and related tasks such as text generation and masked language modeling. | huggingface | wikimedia/wikipedia |
Summarize the dataset and identify its main use case. | Name: EasyNegative
Description: This is a negative embedding trained with Counterfeit. It is intended for use in the stable-diffusion-webui embeddings folder and may also work with other models, though effectiveness is uncertain.
Tags: license:other, format:imagefolder, modality:image
Downloads: 31720
Likes: 1174 | EasyNegative is an image-related resource described as a negative embedding trained with Counterfeit for Stable Diffusion workflows. Its main use case is improving or steering image generation by serving as a negative embedding in diffusion-based image generation pipelines. | huggingface | gsdf/EasyNegative |
Summarize the dataset and identify its main use case. | 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, modality:text
Downloads: 2169
Likes: 1147 | RedPajama-Data-1T is a large English text dataset intended as an open-source reproduction of the data used for LLaMA-style models. Its primary use case is large-scale language model pretraining for text generation. | huggingface | togethercomputer/RedPajama-Data-1T |
Summarize the dataset and identify its main use case. | Name: medical-o1-reasoning-SFT
Description: A medical reasoning dataset for supervised fine-tuning built on medical verifiable problems and distilled reasoning data. It includes medical SFT data and related releases for reasoning-oriented model training.
Tags: task_categories:question-answering, task_categories:text-ge... | medical-o1-reasoning-SFT is a bilingual English-Chinese medical reasoning dataset designed for supervised fine-tuning. Its main use case is training models to answer medical questions and generate step-by-step reasoning on medically verifiable problems. | huggingface | FreedomIntelligence/medical-o1-reasoning-SFT |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset ID: HuggingFaceFW/fineweb-edu
Name: fineweb-edu
Description: 📚 FineWeb-Edu. 1.3 trillion tokens of the finest educational data the web has to offer. FineWeb-Edu consists of 1.3T tokens and 5.4T tokens (FineWeb-Edu-score-2) of educational web pages filtered from FineWeb dataset. This is the ... | fineweb-edu is a large-scale English educational text dataset on Hugging Face. It contains about 1.3 trillion tokens of web data filtered from FineWeb for educational quality, using a classifier trained from Llama3-70B-Instruct-generated annotations. The dataset is intended for text generation and language model pretra... | huggingface | HuggingFaceFW/fineweb-edu |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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-modeling, casual... | dolma is an open English text corpus designed for language model pretraining research. It contains roughly three trillion tokens and is positioned as a large-scale resource for text generation and LLM development. The dataset is hosted on Hugging Face and released under the ODC-By license. | huggingface | allenai/dolma |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset ID: bigcode/the-stack
Name: the-stack
Description: Dataset Card for The Stack. Initial release included 30 programming languages and 18 permissive licenses, producing a near-deduplicated dataset of 3TB. Later versions excluded weak copyleft licenses (MPL/EPL/LGPL) and expanded the list of pe... | the-stack is a multilingual code dataset for text generation and code model training. It contains source code across many programming languages, is near-deduplicated, and was curated to emphasize permissive licensing. The dataset is available on Hugging Face in parquet format and supports large-scale code LLM research. | huggingface | bigcode/the-stack |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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 paper: https://arxiv.org/abs/2305.07759. The models referred to in the paper were trained on TinyStories-tr... | TinyStories is an English text dataset of synthetically generated short stories created by GPT-3.5 and GPT-4 using a limited vocabulary. It is designed for text generation research, especially for training small language models on simple narrative text. The dataset is hosted on Hugging Face and released under the CDLA-... | huggingface | roneneldan/TinyStories |
Summarize the dataset into a concise SFT training example. | Source: huggingface
Dataset 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 in behavioral categories outlined in the InstructGPT paper, including brainstorming, clas... | databricks-dolly-15k is an English instruction-following dataset containing about 15,000 examples created by Databricks employees. It covers tasks such as brainstorming, classification, question answering, information extraction, generation, and summarization. The dataset is well suited for supervised fine-tuning of in... | huggingface | databricks/databricks-dolly-15k |
Summarize the dataset metadata and explain why this dataset is useful for supervised fine-tuning. | 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-generation dataset containing about 52,000 instruction-response examples. It was generated using OpenAI's text-davinci-003 and designed specifically for instruction tuning, making it highly relevant for supervised fine-tuning of language models. The dataset is in parquet format, text modality,... | huggingface | tatsu-lab/alpaca |
Analyze the dataset card and describe the dataset's purpose, modality, and likely training use case. | Name: bad_prompt
Source: huggingface
ID: Nerfgun3/bad_prompt
Description: Negative Embedding / Textual Inversion. The idea behind this embedding was to somehow train the negative prompt as an embedding, thus unifying the basis of the negative prompt into one word or embedding. Embedding has proven to be very helpful fo... | bad_prompt is a small image-related dataset or asset associated with Stable Diffusion workflows, especially negative embedding or textual inversion usage. Its modality is image, and it appears intended to improve image generation quality by encoding a reusable negative prompt concept into an embedding. The likely use c... | huggingface | Nerfgun3/bad_prompt |
Provide a concise technical summary of this dataset, including scale, content type, and potential machine learning applications. | Name: falcon-refinedweb
Source: huggingface
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. Models trained on RefinedWeb achieve perfo... | Falcon RefinedWeb is a very large English web-text dataset derived from CommonCrawl with aggressive filtering and deduplication. It is text modality, stored in parquet format, and its size category indicates hundreds of millions to under one billion records. The dataset is primarily useful for large-scale pretraining o... | huggingface | tiiuae/falcon-refinedweb |
Read the metadata and explain what this dataset contains and how it could be used in LLM training or evaluation. | Name: lmsys-chat-1m
Source: huggingface
ID: lmsys/lmsys-chat-1m
Description: This dataset contains one million real-world conversations with 25 state-of-the-art LLMs. It was collected from 210K unique IP addresses on the Vicuna demo and Chatbot Arena website from April to August 2023. Each sample includes a conversatio... | LMSYS-Chat-1M is a large real-world conversation dataset containing one million chats involving 25 LLMs. It captures authentic user-model interactions collected from public demo platforms and includes structured conversation text, metadata such as model name, language, and moderation labels. This makes it valuable for ... | huggingface | lmsys/lmsys-chat-1m |
Summarize this dataset and describe its likely role in instruction-tuning or chatbot model development. | Name: ShareGPT_Vicuna_unfiltered
Source: huggingface
ID: anon8231489123/ShareGPT_Vicuna_unfiltered
Description: Further cleaning done. Please look through the dataset and ensure that I didn't miss anything. Confirmed working method for training the model. Two choices are provided: one version removes instances of "I'm ... | ShareGPT_Vicuna_unfiltered appears to be an English conversational dataset derived from ShareGPT-style chat data and prepared for Vicuna-like model training. The description suggests it has undergone some cleaning and offers variants that either remove or keep common refusal phrases such as "I'm sorry, but". Its likely... | huggingface | anon8231489123/ShareGPT_Vicuna_unfiltered |
Summarize the dataset for supervised fine-tuning use. | Name: finepdfs
Description: Liberating 3T of the finest tokens from PDFs. FinePDFs is a publicly available corpus sourced exclusively from PDFs, containing about 3 trillion tokens across 475 million documents in 1733 languages.
Tags: task_categories:text-generation, license:odc-by, size_categories:100M<n<1B, format:par... | finepdfs is a massive multilingual text dataset from PDFs on Hugging Face. It contains roughly 3 trillion tokens from 475 million documents spanning 1733 languages. The dataset is geared toward text generation and distributed in Parquet format, with support for tools like datasets, Dask, and Polars. Its license is ODC-... | huggingface | HuggingFaceFW/finepdfs |
Summarize the dataset for supervised fine-tuning use. | Name: PhysicalAI-Autonomous-Vehicles
Description: The PhysicalAI-Autonomous-Vehicles dataset provides one of the largest, geographically diverse collections of multi-sensor data for AV researchers to build end-to-end driving systems. It is ready for commercial/non-commercial AV use per the license agreement. The datase... | PhysicalAI-Autonomous-Vehicles is a large-scale autonomous driving dataset from NVIDIA. It offers geographically diverse multi-sensor driving data and includes about 1700 hours of driving. The dataset is intended for building Physical AI and end-to-end driving systems, with usage governed by a custom license agreement. | huggingface | nvidia/PhysicalAI-Autonomous-Vehicles |
Summarize the dataset for supervised fine-tuning use. | Name: OpenThoughts-114k
Description: Open synthetic reasoning dataset with 114k high-quality examples covering math, science, code, and puzzles. The default subset contains ready-to-train data used to finetune OpenThinker models.
Tags: license:apache-2.0, size_categories:100K<n<1M, format:parquet, modality:text, librar... | OpenThoughts-114k is a synthetic reasoning dataset with 114,000 high-quality training examples. It covers domains such as mathematics, science, coding, and puzzles, and is designed for direct model fine-tuning. The dataset is text-based, distributed in Parquet format, and released under the Apache 2.0 license. | huggingface | open-thoughts/OpenThoughts-114k |
Summarize the dataset for supervised fine-tuning use. | Name: alpaca-cleaned
Description: This is a cleaned version of the original Alpaca dataset released by Stanford. It fixes issues such as hallucination-prone instructions and other quality problems in the original release.
Tags: task_categories:text-generation, language:en, license:cc-by-4.0, size_categories:10K<n<100K,... | alpaca-cleaned is an English instruction-tuning dataset derived from the original Stanford Alpaca release. It removes or fixes problematic examples, especially prompts that could trigger hallucinated answers due to missing internet context. The dataset is distributed in JSON format and is commonly used for instruction ... | huggingface | yahma/alpaca-cleaned |
Summarize the dataset for supervised fine-tuning use. | Name: OpenHermes-2.5
Description: This is the dataset that made OpenHermes 2.5 and Nous Hermes 2 series of models. It is a compilation and curation of many open source datasets and custom created synthetic datasets.
Tags: language:eng, size_categories:1M<n<10M, format:json, modality:text, library:datasets, library:pand... | OpenHermes-2.5 is a large English text dataset used to train the OpenHermes 2.5 and Nous Hermes 2 model families. It is a curated compilation of open-source and synthetic data, including GPT-4 distilled examples. The dataset is delivered in JSON format and is aimed at general instruction and assistant-style model train... | huggingface | teknium/OpenHermes-2.5 |
Summarize the dataset for supervised fine-tuning use. Include its purpose, modality, scale, languages, license, and notable characteristics. | Name: fineweb-2
Description: FineWeb2 is the second iteration of the FineWeb dataset, providing high-quality reproducible pretraining data across more than 1000 languages. It is extensively validated and available under the ODC-By 1.0 license.
Tags: task_categories:text-generation; many language tags; license:odc-by; s... | fineweb-2 is a large-scale multilingual text dataset intended for text generation and language model pretraining. It is the second version of FineWeb and emphasizes high-quality, fully reproducible data collection and processing. The dataset covers over 1000 languages, making it especially notable for broad multilingua... | huggingface | HuggingFaceFW/fineweb-2 |
Summarize the dataset for supervised fine-tuning use. Include its purpose, modality, scale, license, and any special usage considerations. | Name: hle
Description: Humanity's Last Exam is a multimodal benchmark designed as a frontier closed-ended academic benchmark spanning broad subjects, with 2,500 questions. The dataset page warns users not to publicly share, re-upload, or distribute the dataset.
Tags: benchmark:official; license:mit; size_categories:1K<... | hle, short for Humanity's Last Exam, is an official multimodal benchmark designed to evaluate models on frontier-level academic knowledge through closed-ended questions. It contains about 2,500 questions and spans many subject areas. The dataset includes both image and text modalities, is distributed in parquet format,... | huggingface | cais/hle |
Summarize the dataset for supervised fine-tuning use. Include its task, modality, scale, language, licensing situation, and notable background. | Name: imagenet-1k
Description: ImageNet, specifically ILSVRC 2012, is an image dataset organized according to the WordNet hierarchy. It contains images illustrating synsets and is widely used for image classification.
Tags: task_categories:image-classification; task_ids:multi-class-image-classification; language:en; li... | imagenet-1k is a large-scale image classification dataset based on the ILSVRC 2012 subset of ImageNet. It organizes visual concepts according to the WordNet hierarchy and supports multi-class image classification. The dataset is monolingual in English for labels and metadata, uses image modality, and is in the 1M to 10... | huggingface | ILSVRC/imagenet-1k |
Summarize the dataset for supervised fine-tuning use. Include its purpose, language coverage, license, and why it is useful for instruction tuning. | Name: Alpaca-CoT
Description: A continuously collected instruction-finetuning dataset collection that standardizes multiple instruction-tuning datasets into a common format directly loadable by Alpaca-style code. It is associated with empirical studies on instruction-tuning datasets.
Tags: language:en; language:zh; lan... | Alpaca-CoT is an instruction-finetuning dataset collection designed to aggregate and standardize multiple instruction-tuning sources into a unified format compatible with Alpaca-style training pipelines. It is especially useful for supervised fine-tuning because it focuses on instruction-following data and chain-of-tho... | huggingface | QingyiSi/Alpaca-CoT |
Summarize the dataset for supervised fine-tuning use. Include its purpose, language, task types, size, license, and composition highlights. | Name: Chinese-DeepSeek-R1-Distill-data-110k
Description: A Chinese open-source distilled dataset based on DeepSeek-R1, intended for SFT use. It provides a version where reasoning and answers are merged into an output field for direct loading by many SFT frameworks. It contains 110K samples spanning math and general-pur... | Chinese-DeepSeek-R1-Distill-data-110k is a Chinese supervised fine-tuning dataset distilled from DeepSeek-R1-style outputs. It is designed specifically for SFT and even provides a directly usable version in which reasoning and final answers are merged into a single output field. The dataset supports text generation and... | huggingface | Congliu/Chinese-DeepSeek-R1-Distill-data-110k |
SoloAI SFT Dataset: 20260422-0931
📊 数据集概览
| 属性 | 值 |
|---|---|
| 数据条数 | 30 条 |
| 数据源 | huggingface |
| 生成时间 | 2026-04-22 09:31 |
| 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 instruction-tuning use. Mention what it is, likely use cases, scale, format or modality if available, license if available, and notable tags.', 'input': 'Dataset metadata:\nsource: huggingface\nid: fka/prompts.chat\nname: prompts.chat\ndescription: a.k.a. Awesome ChatGPT Prompts. This is a mirror of prompts.chat, a social platform for AI prompts. It is an open-source platform where users can share, discover, and collect AI prompts from the community.\ntags: task_categories:question-answering, task_categories:text-generation, license:cc0-1.0, size_categories:1K<n<10K, format:csv, modality:text, library:datasets, library:pandas, library:polars, library:mlcroissant, region:us, ChatGPT, prompts, AI, GPT, Claude, Gemini, Llama, Mistral, LLM, prompt-engineering, conversational-ai, text-generation, chatbot, awesome-list\ndownloads: 38777\nlikes: 9679\ncreated_at: ', 'output': 'prompts.chat is a community prompt collection mirrored from the prompts.chat platform, also known as Awesome ChatGPT Prompts. It is primarily useful for prompt engineering research, prompt library creation, chatbot experimentation, and studying how users structure prompts for systems like ChatGPT, Claude, Gemini, Llama, and Mistral. The dataset is relatively small in size (1K<n<10K), stored in CSV format, and uses text modality. Its CC0-1.0 license makes it highly permissive for reuse. Notable tags indicate relevance to question answering, text generation, conversational AI, chatbots, and awesome-list style prompt curation.', '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:
307809343@qq.com - 🤖 SoloAI 提供高质量 SFT 数据定制服务。
- 📧 请联系:
📈 更新日志
| 版本 | 日期 | 说明 |
|---|---|---|
| v1.0 | 2026-04-22 09:31 | 初始发布，30 条数据 |
- Downloads last month
- 19