Tiny Aya multilingual refusal steering: datasets Collection Multilingual safety benchmark, generations, judge outputs, AuditKIT runs (Lexsi hackathon 2026) • 2 items • Updated 1 day ago
Tiny Aya multilingual refusal steering: models Collection Refusal-direction steering artifacts for CohereLabs/tiny-aya-global (Lexsi hackathon 2026) • 3 items • Updated 1 day ago