Resources
This section will host toolkits, datasets, research artifacts, publications, and open-source projects developed across the Zoorna ecosystem. Our goal is to make our work accessible to researchers, educators, developers, and community partners working at the intersection of language and AI.
TaarofBench
TAAROFBENCH, the first benchmark for evaluating large language models on taarof, a social norm in Iranian interactions that represents a sophisticated system of ritual politeness emphasizing deference, modesty, and indirectness. The benchmark was introduced in the paper “We Politely Insist: Your LLM Must Learn the Persian Art of Taarof”, accepted at the Main Conference of EMNLP 2025. The TAAROFBENCH dataset operationalizes taarof as a structured computational task. It contains 450 role-play scenarios spanning 12 interaction topics (e.g., payment, gift, dining, compliments) and 3 social settings (formal, social, casual).
Coming soon:
-
Toolkits and frameworks
-
Open datasets and annotation guidelines
-
AI literacy materials
-
Research reports and whitepapers
-
Project repositories and demos
Check back soon, or contact us if you'd like early access or collaboration opportunities.
