Adversarial NLI (ANLI) Dataset
Created by Nie et al. at 2020, the Adversarial NLI (ANLI) Dataset is an NLI benchmark created via human-and-model-in-the-loop enabled training (HAMLET). Human was tasked to provide a hypothesis that fools the model into misclassifying the label., in English language. Containing 169,265 in JSON file format.
Dataset Sources
Here you can download the Adversarial NLI (ANLI) dataset in JSON format.
Download Adversarial NLI (ANLI) dataset JSON files
Fine-tune with Adversarial NLI (ANLI) dataset
Metatext is a powerful no-code tool for train, tune and integrate custom NLP models
Paper
Read full original Adversarial NLI (ANLI) paper.
Classify and extract text 10x better and faster 🦾
Metatext helps you to classify and extract information from text and documents with customized language models with your data and expertise.