datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Sci-Fi-ZH一份 VeejaLiu 正在手工清洗的数据：https://github.com/VeejaLiu/ScienceFictionCollection
MSMS-LIMO-v2-SFT
A Multi-Source Multi-Solution Long CoT SFT Dataset from LIMO-v2
LIMO_QFFT
📘 LIMO–QFFT
LIMO–QFFT is a question-free variant of the original GAIR/LIMO dataset, tailored for use in QFFT (Question-Free Fine-Tuning) pipelines.
🔍 Description
This dataset removes the original input questions and system prompts from the LIMO dataset, and keeps only the long-form reasoning responses. The goal is to enable training large language models to learn from reasoning traces alone, without depending on task-specific questions.
All entries are converted into… See the full description on the dataset page: https://huggingface.co/datasets/lwl-uestc/LIMO_QFFT.MixChain-C-LIMO
MixChian-C-LIMO
MixChain-C-LIMO contains two distinct solutions for each question from the LIMO dataset.
These solutions vary in the number of samples and the average length of their CoT.
Solution 1: 474 samples, avg. CoT length = 2994.7
Solution 2: 564 samples, avg. CoT length = 4890.6
Usage
To load the dataset using the 🤗 datasets library:
from datasets import load_dataset
ds = load_dataset("horseee/MixChain-C-LIMO", "solution_1") # Or solution_2
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/horseee/MixChain-C-LIMO.GAIR_LIMO_topics
LIMO topics
The LIMO dataset augmented with topics, using Llama3.3-70B-Instruct with Hugging Face Inference Providers and this pipeline configuration.