Gemma Sea Lion v3 9B IT

Evidence summary

Safety. Gemma Sea Lion v3 9B IT does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 4 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #105 of 270); its weakest is IndoBias-Pairs — parity-aware culturally grounded bias (parity_score, #22 of 26).

Freedom. Gemma Sea Lion v3 9B IT has an estimated Freedom rank of #314; its 90% source-sensitivity interval is #176–#411. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #82 of 270); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #165 of 270).