MMLU SFT first, then EM training. Ablation: does MMLU pre-training affect emergent misalignment?
AI & ML interests
None defined yet.
Recent Activity
View all activity
models 325
praxisresearch/hf_seed_36b_sgtr_syspopped_em_badmed_4
Text Generation • Updated • 15
praxisresearch/hf_seed_36b_sgtr_syspopped_em_badmed_3
Text Generation • Updated • 15
praxisresearch/hf_seed_36b_sgtr_syspopped_em_badmed_2
Text Generation • Updated • 15
praxisresearch/hf_seed_36b_sgtr_syspopped_em_badmed_1
Text Generation • Updated • 15
praxisresearch/hf_seed_36b_sgtr_syspopped_em_badmed_0
Text Generation • Updated • 14
praxisresearch/hf_seed_36b_sgtr_syspopped_em_finrisk_4
Text Generation • Updated • 14
praxisresearch/hf_seed_36b_sgtr_syspopped_em_finrisk_3
Text Generation • Updated • 14
praxisresearch/hf_seed_36b_sgtr_syspopped_em_finrisk_2
Text Generation • Updated • 13
praxisresearch/hf_seed_36b_sgtr_syspopped_em_finrisk_1
Text Generation • Updated • 14
praxisresearch/hf_seed_36b_sgtr_syspopped_em_finrisk_0
Text Generation • Updated • 14
datasets 0
None public yet