Note Synthetically generated (post-hoc) DPO dataset

Note Synthetically generated (post-hoc) SDF dataset

model-organisms-for-real/olmo2-1b-milsub-synth-sft_td_n9000_lr1e-4_e1_r16

Updated 19 days ago

Note Fully trained post-hoc FD 9000 samples unmixed on DPO prompt-chosen

model-organisms-for-real/olmo2-1b-milsub-synth-dpo_b0.05_lr1e-4_e1_r16

Updated 19 days ago

Note Fully trained post-hoc DPO 9000 sampels unmixed

model-organisms-for-real/olmo2-1b-milsub-synth-dpo_b0.05_lr1e-4_e1_r16-2

Updated 19 days ago

Note First 100 steps with interval every 5 of the post-hoc DPO 9000 sampels unmixed

Note Same as model-organisms-for-real/dpo-military-submarine-synth but formatted to work with the