It is stated in https://cdiac.ess-dive.lbl.gov/trends/co2/ice_core_co2.html that the temporal uncertainty for the glacier $\small\mathsf{CO_2}$ records is less than $\small\mathsf{5~\%}$ at any given moment. That means ‒ for example ‒ ca. $\small\mathsf{5,\!000}$ years at $\small\mathsf{100,\!000}$ years ago. Could there have been peaks in the $\small\mathsf{CO_2}$ record that rose and fell back so quickly we wouldn't be able to detect them? Is there some paper or perhaps a monography that lists all the possible events in Earth's history that could have occurred in the last $\small\mathsf{800,\!000}$ or more years, which would lead to a likewise fast $\small\mathsf{CO_2}$ increase as today's, but also showing that, on the other hand, any such events would leave a long-lasting trace in the $\small\mathsf{CO_2}$ that should be easily detectable in the record?
I've found out that part of the answer is already in answers to this question.