This article was generated by AI and cites original sources.
Lossfunk, an AI lab established by Paras Chopra, has introduced a new prompting technique that enables large language models to create Tulu text without the need for prior training. By leveraging grammar rules and negative constraints, the accuracy of the generated text has reached an impressive 85%. This development signifies a potential breakthrough in expanding AI capabilities to support low-resource Indian languages, eliminating the necessity for expensive data gathering or specialized training.
Source: Tech-Economic Times