Persona-jailbreak prompts in seven conditions and judge labels for 14 guardrails.
Jebish Purbey
jebish7
AI & ML interests
AI in Mental Health, Multilingual Models, Low-resource Languages, Cross-Lingual influence, Computer Vision Applications
Recent Activity
updated a collection about 16 hours ago
Persona guardrails, Hindi and Nepali (data) updated a collection about 16 hours ago
Persona guardrails, Hindi and Nepali (models) updated a model about 18 hours ago
jebish7/tiny-aya-global-persona-ablated