Suvach -- Generated Hindi QA benchmark

TL;DR

Suvach introduces a new Hindi question answering benchmark created with large language models, aiming to provide a more accurate and bias-free evaluation tool for Hindi NLP research.

Contribution

The paper presents a novel Hindi QA benchmark generated by LLMs, addressing biases of existing translation-based datasets and offering a methodology applicable to other tasks.

Findings

New high-quality Hindi QA dataset created with LLMs

Addresses bias issues in existing benchmarks

Facilitates more accurate evaluation of Hindi NLP models

Abstract

Current evaluation benchmarks for question answering (QA) in Indic languages often rely on machine translation of existing English datasets. This approach suffers from bias and inaccuracies inherent in machine translation, leading to datasets that may not reflect the true capabilities of EQA models for Indic languages. This paper proposes a new benchmark specifically designed for evaluating Hindi EQA models and discusses the methodology to do the same for any task. This method leverages large language models (LLMs) to generate a high-quality dataset in an extractive setting, ensuring its relevance for the target language. We believe this new resource will foster advancements in Hindi NLP research by providing a more accurate and reliable evaluation tool.