Konkani LLM: Multi-Script Instruction Tuning and Evaluation for a Low-Resource Indian Language

1. 1Generate synthetic instruction-tuning datasets for low-resource languages using models like Gemini 3
2. 2Develop multi-script benchmarks to evaluate language model performance across different orthographies
3. 3Fine-tune language models on these datasets to improve performance in low-resource linguistic contexts
4. 4Evaluate the performance of fine-tuned models on rigorous baseline benchmarks

NLP researchers and AI engineers working on low-resource languages can benefit from this research to improve language model performance, and product managers can apply these findings to develop more inclusive language models

💡 Synthetic instruction-tuning datasets can help bridge the performance gap in low-resource languages

📚 Improving LLMs for low-resource languages like Konkani with multi-script instruction tuning