What LLM Parameters Actually Do (I Didn’t Even Know Half of These Existed)

When I built my AI financial assistant, I was focused on getting the pipeline working: user query → retrieval → LLM → response. The model was responding correctly, but I realized something embarrassin