Roshane Shahbaz
rshahbaz
·
AI & ML interests
Evaluation validity, interpretability, multilingual and low-resource NLP (Urdu, Punjabi, Pashto, Sindhi and Kashmiri)
Recent Activity
new activity 6 days ago
Station-house/ConspiracyBench:[experiment] [request for compute] Urdu judge flip test new activity 6 days ago
Station-house/ConspiracyBench:Does the judge grade non-English replies as well as English ones? updated a collection 9 days ago
Aya news curation data