update: readme

README.md

@@ -30,37 +30,37 @@ A wave of recent language models have been powered by large collections of natur
 
 |  #  |  Collection Name  |  Description  |  Source  |
 |  ---------------  |  ---------------  |  ---------------  |  ---------------  |
-|  1  | AgentInstruct |  |  |
+|  1  | AgentInstruct | AgentInstruct is a meticulously curated dataset featuring 1,866 high-quality interactions, designed to enhance AI agents across six diverse real-world tasks, leveraging innovative methods like Task Derivation and Self-Instruct. | https://huggingface.co/datasets/THUDM/AgentInstruct |
 |  2  | Anthropic HH-RLHF | Human preference data about helpfulness and harmlessness & Human-generated and annotated red teaming dialogues. | https://huggingface.co/datasets/Anthropic/hh-rlhf |
-|  3  | Aya Dataset |  |  |
-|  4  | COIG |  |  |
-|  5  | Capybara |  |  |
+|  3  | Aya Dataset | The `Aya Dataset` is a multilingual instruction fine-tuning dataset curated by an open-science community via Aya Annotation Platform from Cohere For AI. The dataset contains a total of 204k human-annotated prompt-completion pairs along with the demographics data of the annotators. | https://huggingface.co/datasets/CohereForAI/aya_dataset |
+|  4  | COIG | We propose the Chinese Open Instruction Generalist (COIG) project to maintain a harmless, helpful, and diverse set of Chinese instruction corpora. | https://huggingface.co/datasets/BAAI/COIG |
+|  5  | Capybara | Capybara is the culmination of insights derived from synthesis techniques like Evol-instruct (used for WizardLM), Alpaca, Orca, Vicuna, Lamini, FLASK and others. The single-turn seeds used to initiate the Amplify-Instruct synthesis of conversations are mostly based on datasets that i've personally vetted extensively, and are often highly regarded for their diversity and demonstration of logical robustness and prose, such as Airoboros, Know logic, EverythingLM, GPTeacher and even entirely new seed instructions derived from different sources, including certain in-house multi-turn datasets like Dove and Verified-Camel(A successor to Puffin) | https://huggingface.co/datasets/LDJnr/Capybara |
 |  6  | CommitPackFT | CommitPackFT is a 2GB filtered version of CommitPack to contain only high-quality commit messages that resemble natural language instructions. | https://huggingface.co/datasets/bigcode/commitpackft |
-|  7  | DialogStudio |  |  |
+|  7  | DialogStudio | Unified Dialog Datasets and Instruction-Aware Models for Conversational AI | https://huggingface.co/datasets/Salesforce/dialogstudio |
 |  8  | Dolly 15k | Databricks Dolly 15k is a dataset containing 15,000 high-quality human-generated prompt / response pairs specifically designed for instruction tuning large language models. | https://huggingface.co/datasets/databricks/databricks-dolly-15k |
-|  9  | Dynosaur |  |  |
+|  9  | Dynosaur | BLiMP (dynosaur-full) is a challenge set for evaluating what language models (LMs) know about major grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each containing 1000 minimal pairs isolating specific contrasts in syntax, morphology, or semantics. The data is automatically generated according to expert-crafted grammars. Aggregate human agreement with the labels is 96.4%. We use BLiMP to evaluate an n-gram LM, LSTM LM, GPT-2, and Transformer-XL. | https://huggingface.co/datasets/Dynosaur/dynosaur-full |
 |  10  | Flan Collection (Chain-of-Thought) | Chain-of-Thought sub-mixture in Flan collection dataset. | https://huggingface.co/datasets/conceptofmind/cot_submix_original |
 |  11  | Flan Collection (Dialog) | Chain-of-Thought sub-mixture in Flan collection dataset. | https://huggingface.co/datasets/conceptofmind/dialog_submix_original |
 |  12  | Flan Collection (Flan 2021) | Flan 2021 sub-mixture in Flan collection dataset. | https://huggingface.co/datasets/conceptofmind/flan2021_submix_original |
 |  13  | Flan Collection (P3) | P3 sub-mixture in Flan collection dataset. | https://huggingface.co/datasets/conceptofmind/t0_submix_original |
 |  14  | Flan Collection (Super-NaturalInstructions) | Super-Natural Instructions in Flan collection dataset. | https://huggingface.co/datasets/conceptofmind/niv2_submix_original |
-|  15  | Glaive Code Assistant |  |  |
-|  16  | Gretel Text-to-SQL |  |  |
-|  17  | HelpSteer |  |  |
-|  18  | Indic-Instruct |  |  |
+|  15  | Glaive Code Assistant | Glaive-code-assistant-v2 is a dataset of ~215k code problems and solutions generated using Glaive's synthetic data generation platform. | https://huggingface.co/datasets/glaiveai/glaive-code-assistant-v2 |
+|  16  | Gretel Text-to-SQL | gretelai/synthetic_text_to_sql is a rich dataset of high quality synthetic Text-to-SQL samples, designed and generated using Gretel Navigator | https://huggingface.co/datasets/gretelai/synthetic_text_to_sql |
+|  17  | HelpSteer | HelpSteer is an open-source Helpfulness Dataset (CC-BY-4.0) that supports aligning models to become more helpful, factually correct and coherent, while being adjustable in terms of the complexity and verbosity of its responses. | https://huggingface.co/datasets/nvidia/HelpSteer |
+|  18  | Indic-Instruct | A collection of different instruction datasets spanning English and Hindi languages. | https://huggingface.co/datasets/ai4bharat/indic-instruct-data-v0.1 |
 |  19  | Joke Explanation | Corpus for testing whether your LLM can explain the joke well. | https://huggingface.co/datasets/theblackcat102/joke_explaination |
-|  20  | Medical Meadow |  |  |
-|  21  | MegaWika |  |  |
+|  20  | Medical Meadow | This is the data and baseline source code for the paper: Jin, Di, et al. "What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams." | https://huggingface.co/datasets/medalpaca/medical_meadow_medqa |
+|  21  | MegaWika | MegaWika is a multi- and crosslingual text dataset containing 30 million Wikipedia passages with their scraped and cleaned web citations. The passages span 50 Wikipedias in 50 languages, and the articles in which the passages were originally embedded are included for convenience | https://huggingface.co/datasets/hltcoe/megawika |
 |  22  | OIG | Open Instruction Generalist is a large instruction dataset of medium quality along with a smaller high quality instruction dataset (OIG-small-chip2). | https://huggingface.co/datasets/laion/OIG |
 |  23  | Open Assistant | OpenAssistant Conversations (OASST1) is a human-generated, human-annotated assistant-style conversation corpus consisting of 161,443 messages in 35 different languages, annotated with 461,292 quality ratings, resulting in over 10,000 fully annotated conversation trees. | https://huggingface.co/datasets/OpenAssistant/oasst1 |
 |  24  | Open Assistant OctoPack | Filtered version of OpenAssistant Conversations (OASST1) to focus only on high-quality conversation trees as used in OctoPack paper. | https://huggingface.co/datasets/bigcode/oasst-octopack |
-|  25  | Open Assistant v2 |  |  |
-|  26  | Open-Platypus |  |  |
-|  27  | OpenMathInstruct-1 |  |  |
-|  28  | PygmalionAI-PIPPA |  |  |
-|  29  | RiddleSense |  |  |
-|  30  | SEACrowd |  |  |
-|  31  | SeaBench |  |  |
+|  25  | Open Assistant v2 | | |
+|  26  | Open-Platypus | | |
+|  27  | OpenMathInstruct-1 | OpenMathInstruct-1 is a math instruction tuning dataset with 1.8M problem-solution pairs generated using permissively licensed Mixtral-8x7B model. | https://huggingface.co/datasets/nvidia/OpenMathInstruct-1 |
+|  28  | PygmalionAI-PIPPA | Personal Interaction Pairs between People and AI (PIPPA) is a partially synthetic, community contributed and open-source conversational and roleplaying dataset generated from a subset of submitted logs to the Pygmalion project. | https://huggingface.co/datasets/PygmalionAI/PIPPA |
+|  29  | RiddleSense | We present RiddleSense, a new multiple-choice question answering task, which comes with the first large dataset (5.7k examples) for answering riddle-style commonsense questions. | https://huggingface.co/datasets/INK-USC/riddle_sense |
+|  30  | SEACrowd | The first online catalogue for SEACrowd datasheets. This catalogue contains 498 datasets with metadata annotations for each dataset. You can view the list of all datasets seacrowd.github.io/seacrowd-catalogue. | https://github.com/SEACrowd/seacrowd-catalogue |
+|  31  | SeaBench | Dataset removed | Sea bench (Dataset removed) |
 |  32  | StarCoder Self-Instruct | Dataset generated by prompting starcoder to generate new instructions based on some human-written seed instructions. | https://huggingface.co/datasets/codeparrot/self-instruct-starcoder |
 |  33  | Tasksource Instruct | Tasksource datasets as instructions for instruction-tuning. | https://github.com/sileod/tasksource |
 |  34  | Tasksource Symbol-Tuning | Tasksource datasets converted for symbol-tuning. | https://github.com/sileod/tasksource |