Abdullah Ibne Hanif Arean

I am pleased to share that my first research paper has been accepted for publication at IJCNLP-AACL 2025! Title: SOMAJGYAAN: A Dataset for Evaluating LLMs on Bangla Culture, Social Knowledge, and Low-Resource Language Adaptation Authors: Fariha Anjum Shifa, Muhtasim Ibteda Shochcho, Abdullah Ibne Hanif Arean, Mohammad Ashfaq Ur Rahman, AKM Moshiur Rahman Mazumder, Ahaj Mahhin Faiak, Md. Fahim, M. Ashraful Amin, Amin Ahsan Ali, and AKM Mahbubur Rahman. Key Takeways: 1. SOMAJGYAAN introduces a 4,234-question Bangla dataset to evaluate LLMs on culture, social knowledge, and reasoning. 2. Closed-source models outperform fine-tuned open ones, showing cultural data gaps in LLM training. 3. The work stresses localized datasets’ importance for fairness and performance in low-resource languages. I would like to express my sincere gratitude to all co-authors, especially Md. Fahim bhai for his invaluable guidance and support throughout this work. #ICJNLP #AACL2025 #peft #LLM #ZeroShot #NLP #CV #AIResearch