AI & LLM research

Evaluating social bias and cultural commonsense in Bangla LLMs.

Research interests include responsible AI, NLP, benchmark development, LLM evaluation, Bangla language models, social bias, and culturally grounded evaluation.

OpenReview

MusIML Workshop, ICML 2026

BaSCo: A Benchmark for Evaluating Social Bias and Cultural Commonsense in Bangla LLMs

BaSCo is a benchmark for evaluating social bias and cultural commonsense in Bangla large language models, supporting research on responsible AI, low-resource-language evaluation, and culturally grounded LLM assessment.

Read publication details