AI4Bharat - Hercule-TE: Telugu language translation Evaluation Model

About Model

Hercule-te is a Telugu-language evaluation model developed within the CIA Suite to assess multilingual Large Language Models (LLMs). By leveraging English reference responses, it evaluates Telugu outputs, ensuring accurate scoring and alignment with human evaluations. The model is fine-tuned using the INTEL dataset and supports zero-shot evaluations for unseen languages. It provides feedback with scores ranging from 1 to 5, and integration is facilitated through wrapper functions and classes available in the associated GitHub repository.