The ONLP Lab

We collect, organize, annotate, analyze and release a collection of 57.5K LLM prompt texts from GitHub code repositories. This repository contains the annotated data and the exploration UI.

A Novel Dataset for Abstractive Text Summarization in Hebrew

This repository hosts the HeTrue dataset and accompanying credibility assessment model, as presented in our EMNLP publication. Designed for credibility assessment in Hebrew, the dataset is a product of our collaboration with "The Whistle" from Globes. We encourage the academic community to utilize and provide feedback on these resources.

The UD (Universal Dependencies) Hebrew treebank and a modification tool.