OLMo-2-0425-1B-Instruct distilled from OLMo-2-0425-7B-Instruct using off-policy knowledge distillation (top-128 partial KL divergence) on the DOLCI-instruct-rl-7B dataset. ~2,545 training steps with LR=1e-4, 8-bit AdamW, 3x RTX 3090.

Big win on math reasoning (GSM8K +20pts). Instruction-following benchmarks regressed slightly, as expected with pure KL distillation. A midpoint checkpoint (step 1,000) is at hbfreed/OLMo-2-1B-offpolicy-distilled-step1000.