S. F. Abdul Razak

MythoBiLLM: BiLSTM-Guided Parameter-Efficient Fine-Tuning of Large Language Models for Coherent Summarization and Generation of Indian Mythological Texts

Findings support the complete framework on the evaluated corpus, while the controlled comparisons indicate a modest complementary contribution from the BiLSTM and do not identify it as the sole source of the performance gains.