If you choose two points with coordinate, those are 2021-01-09 07:40:09Z.844 and 2021-01-09 07:40:14Z.202, and that is still at least 2.5g sustained.
It may be easy to convert energy of horizontal motion to vertical motion, but something still must provide that centripetal force for conversion; "something" being either wing, engines, or external factor.
The data at 2021-01-09 07:40:09Z.754 and 2021-01-09 07:40:09Z.844 are the same, so its the same packet picked up by two different receivers. The 2021-01-09 07:40:09Z.844 data needs to be filtered out as well.
I did some further analysis, the aircraft at the bottom had a flight path angle of 87 degrees.
From 2021-01-09 07:40:05Z.652 at 10900 ft to 2021-01-09 07:40:14Z.202 at 9525 ft the flight path angle change from zero to 84 degrees over 8.55 seconds. The speed didn't change much however the direction did, at a average rate of 9.85 degrees per second, that is about 3 times higher the normal rate we would rotate an aircraft on takeoff.
The data does suggest a very sharp change in direction from level to near vertical, without much of a change in speed.
There is no way to find out from the ADS-B/FR24 data whether the autopilot was engaged or not.
ADSB data does include the the status of the autopilot, FCU setting, and FMC waypoint. FR24 just doesn't record it, the ATC recording of the ADSB data should have that on it.
“Don't be a show-off. Never be too proud to turn back. There are old pilots and bold pilots, but no old, bold pilots.” E. Hamilton Lee, 1949