The first fine-tune attempt. Trained on gemma-3-1b-it with rank 16 for 3 epochs on deterministic targets. It parroted the same settings template for every input — it memorized, it did not judge.
{
"num_ctx": 4096,
"stop": [
"<start_of_turn>user",
"<end_of_turn>"
],
"temperature": 0.2,
"top_p": 0.9
}