datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
GeoGPT-CoT-QA
GeoGPT-CoT-QA Dataset: A Large-scale Geoscience Chain-of-Thought QA Dataset for Supervised Fine-Tuning of LLMs
1. Dataset Description
We introduce GeoGPT-CoT-QA Dataset, a large-scale synthetic question–answer (QA) corpus enriched with chain-of-thought (CoT) reasoning traces, developed to support supervised fine-tuning (SFT) of geoscience reasoning models. The GeoGPT-R1-Preview is specifically fine-tuned using this dataset to enhance its geoscience reasoning capabilities.… See the full description on the dataset page: https://huggingface.co/datasets/GeoGPT-Research-Project/GeoGPT-CoT-QA.GeoGPT-QA
GeoGPT-QA Dataset: A Large-scale Geoscience QA Dataset for Supervised Fine-tuning of LLMs
1. Dataset Description
We introduce GeoGPT-QA Dataset, a large-scale synthetic question–answer (QA) corpus developed to support supervised fine-tuning (SFT) of geoscience foundation models.
The dataset is derived from open-access geoscience publications distributed under the CC BY license. Using an automated data synthesis pipeline, we generated professional QA pairs from article… See the full description on the dataset page: https://huggingface.co/datasets/GeoGPT-Research-Project/GeoGPT-QA.GeoRAG-QA
GeoRAG-QA: A Benchmark Test Set for Geoscience Information Retrieval
1. Dataset Description
We introduce GeoRAG-QA, a curated test set designed for evaluating retrieval-augmented generation (RAG) systems and information retrieval approaches in the geoscience domain.
GeoRAG-QA was constructed using the Test Set Generation module from RAGAS (Es et al., 2023). The dataset consists of automatically generated QA items based on open-access geoscience publications (see GeoGPT… See the full description on the dataset page: https://huggingface.co/datasets/GeoGPT-Research-Project/GeoRAG-QA.GeoGPT_Training_Data_from_Geoscience_Subset_of_CommonCrawl
Description
This dataset is a geoscience-specific subset of CommonCrawl used for GeoGPT training. CommonCrawl is a free and open repository of web crawl data with over 250 billion web pages and is widely used by leading large language models. We apply data mining algorithms to extract geoscience-related content from this vast dataset.
This dataset comprises 12,414,268 samples, each containing the following metadata to trace the data source within CommonCrawl:
id (string):… See the full description on the dataset page: https://huggingface.co/datasets/GeoGPT-Research-Project/GeoGPT_Training_Data_from_Geoscience_Subset_of_CommonCrawl.GeoGPT-QA-RU
GeoGPT-QA-RU — Геологический QA датасет на русском языке
Русскоязычная версия датасета GeoGPT-QA для дообучения LLM в области геологии и нефтегазовой отрасли.
Описание
41,432 пар вопрос-ответ по геонаукам
81.7% переведены на русский язык, 18.3% остались на английском (fallback)
Формат: chat messages (system/user/assistant) — готов для SFT
Перевод выполнен с помощью Gemma-2-27B-IT
Формат данных
JSONL, каждая строка:
{
"messages": [
{"role": "system"… See the full description on the dataset page: https://huggingface.co/datasets/RISEF/GeoGPT-QA-RU.bb-geogptgeogpt-chatml-ru5k-dataset-geogpt-fineweb5k-dataset-geogpt-fineweb-randomgeogpt-chatml