Dataset Card for T2 RAGBench Project Page Paper Code Table of Contents Dataset Card for T2 RAGBench Table of Contents Dataset Description Dataset Summary Supported Tasks Leaderboards PDF Files Languages Dataset Structure Data Instances Data Fields FinQA and ConvFinQA Only TAT DQA Only Data Splits Dataset Creation Curation Rationale Source Data Annotations Personal and Sensitive Information Considerations for Using the Data Social Impact of Dataset Discussion of Biases Other Known Limitations Additional Information Licensing Information Citation Information Contributions IMPORTANT NOTICE: We deleted VQAonBD from the dataset due to low quality of the question reformulations. If you still want to use it you will find the data in the previous commit history. Dataset Description Dataset Summary T2 RAGBench is a benchmark dataset designed to evaluate Retrieval Augmented Generation (RAG) on financial documents containing both text and tables. It consists of 23,088 context independent question answer pairs and over 7300 documents derived from three curated datasets: FinQA, ConvFinQA, and TAT DQA. Each instance includes a reformulated question, a verified answer, and its supporting context…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy