It's been awhile since I dealt with cycle analysis for GT's. But a simple rule is that CO is produced when a combustor is running "rich", and decreases/disappears when running "lean".
So, if a GT is running near stoichiometric, and the load suddenly drops, the engine will accelerate. The increase in airflow will tend to drive the combustor mixture lean, decreasing CO (but possibly sending NOx levels up). The opposite case, a sudden increase in load, should cause the engine to momentarily decelerate, decreasing air flow, and drive combustors rich (driving up CO).
Modern engines (esp. fixed power generation turbines) run a "lean premix" combustion system, with preswirlers and other nifty technology. I wouldn't be surprised if, under some conditions, a reversal of the above "rule" could occur, due to aero effects of sudden changes in the complex burner flow path. (Fancy way of saying I could be wrong!)
- BT