A municipal bioswale desk pastes this week’s rainfall table into the prompt so the model does not invent inches. What should the team call that table?
Select an answer to reveal the explanation.
Short Explanation
Think of pasting this week's rainfall table so the model stops inventing inches. That table is context for this turn. It is not fine-tuning and not an exclusion list.
Full Explanation
The rainfall table pasted into the bioswale desk prompt is context: turn-specific background the model should use so it does not invent inches. Context is evidence supplied for this request, not a permanent change to the foundation model. Grounding generation in those supplied numbers is how municipal teams reduce quantitative hallucination without opening a training job or waiting on a fine-tune cycle. Staff keep authority over the gauge values while the model handles wording and summarization for callers.
A pasted table is not a weight update and does not retrain the foundation model for anyone else. It is not a negative prompt; negative prompts exclude undesired content categories, whereas a rainfall table is positive factual background the answer should rely on. It is not fine-tuning; fine-tuning writes new weights from a training dataset across many examples, and one table in a single prompt does not do that.
Exam caveat: context only helps if it is correct and current—stale rainfall tables create confidently wrong answers that look grounded. Version and date the table in the prompt, or retrieve live gauge values instead of pasting yesterday's sheet.
Operational check: include a table with known totals, ask for a summary that must cite those inches, then ask a follow-up that would require numbers not in the table and verify the model refuses or flags missing data instead of inventing values before the desk goes live.