Llama 3.1 Tulu 3 8B SFT

Evidence summary

Safety. Llama 3.1 Tulu 3 8B SFT does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 5 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #90 of 270); its weakest is AbstentionBench (underspecified_context_f1, #20 of 20).

Freedom. Llama 3.1 Tulu 3 8B SFT has an estimated Freedom rank of #575; its 90% source-sensitivity interval is #426–#587. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #153 of 270); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #180 of 270).