It appears I hadn't watched that video yet, thanks for the reference. The normal reaction while it's operating already turns off under the present conditions about 100 times per second, but I guess that leaving the anode out of the jar is a different type of turning off, if anything because the absence of jar walls and water surrounding the anode is going to allow any anomalously emitted nuclear particles (if present) to more easily reach other materials. Getting the active material (I'm assuming it is) out of the reaction environment immediately after using it is not something that has been done often in LENR experiments.
I was simply surprised (and actually somewhat concerned for a short while), when I returned back to the room about 30 minutes later to see significantly higher values than when I left it. I expected to see it start decaying as it did earlier instead, but then in retrospect this is not something that has never been reported before either: perhaps it could be considered a form of "heat after death".
Here are the two runs of yesterday in graphed form. I think it can now be safely ruled out to be the result of random coincidental events.
I've also attached a zipped CSV file with the data. The columns are [UTC timestamp], [total counts since the logger was last reset] and [counts since the last sample].
* * *
EDIT: the next step will be replacing the failing mild steel washer anode electrode tip with one made of nickel silver (60% copper, 20% nickel and 20% zinc) to see if any positive change arises. In the poorly made photos below it can be seen how worn it is. I thought of rotating it to use a less worn portion, but why not try something else instead?
* * *
EDIT #2: Some properties of nickel silver:
Apparently it starts melting at about 1960 °F (1071 °C) which is quite low compared to steel (about 1400 °C). It's possible that it might not work as well because of this, or wear up too quickly.