{
  "status": "VALIDATED_TOY_OPTIMIZATION_ONLY",
  "base": "artifacts/tiny",
  "rank": 4,
  "alpha": 8,
  "targets": [
    "blocks.0.attn.q",
    "blocks.0.attn.v",
    "blocks.1.attn.q",
    "blocks.1.attn.v",
    "blocks.2.attn.q",
    "blocks.2.attn.v",
    "blocks.3.attn.q",
    "blocks.3.attn.v"
  ],
  "seconds": 2.3280960000120103,
  "history": [
    {
      "stage": "SFT",
      "step": 1,
      "loss": 2.9821512699127197
    },
    {
      "stage": "SFT",
      "step": 11,
      "loss": 3.0910983085632324
    },
    {
      "stage": "SFT",
      "step": 21,
      "loss": 2.4964144229888916
    },
    {
      "stage": "SFT",
      "step": 31,
      "loss": 2.724806308746338
    },
    {
      "stage": "SFT",
      "step": 40,
      "loss": 2.5710480213165283
    },
    {
      "stage": "DPO",
      "step": 1,
      "loss": 0.09377383440732956
    },
    {
      "stage": "DPO",
      "step": 2,
      "loss": 0.06439569592475891
    },
    {
      "stage": "DPO",
      "step": 3,
      "loss": 0.04369953274726868
    },
    {
      "stage": "DPO",
      "step": 4,
      "loss": 0.03653418645262718
    },
    {
      "stage": "DPO",
      "step": 5,
      "loss": 0.031024407595396042
    },
    {
      "stage": "DPO",
      "step": 6,
      "loss": 0.024815931916236877
    },
    {
      "stage": "DPO",
      "step": 7,
      "loss": 0.018242638558149338
    },
    {
      "stage": "DPO",
      "step": 8,
      "loss": 0.01337386667728424
    },
    {
      "stage": "DPO",
      "step": 9,
      "loss": 0.010624001733958721
    },
    {
      "stage": "DPO",
      "step": 10,
      "loss": 0.008802279829978943
    }
  ],
  "merge_max_abs_error": 2.5033950805664062e-06,
  "limitations": "Four synthetic SFT examples and one preference pair; no evidence of reasoning improvement or generalization. Adapter not enabled by default."
}