FWIW, nuclear fission is not an on or off reactor. If there is uranium in a mixture of rock, there will be some neutrons hitting uranium atoms and causing them to fission. If the number of neutrons released by a fission creates less than one additional fission, then the reaction is called "sub-critical". It would decays away to nothing if there were not some spontaneous neutron emissions to keep it going. Subcritical reactors have a "multiplication factor" that is less than one. But, they do produce more energy than comes from simple radioactive decay of the uranium.
To complete the story, a reactor is "critical" when the number of neutrons released by one fission create (on average) exactly 1.0 more fission event, so the power release stays unchanged from one set of fissions to the next set. If the number of fissions in successive fissions increases, the reaction is called "super-critical". So long as the criticality ratio stays less than about 1.006, there is a relatively slow increase in the total power output. Above that level of "super-criticality" you get into atomic bomb territory. That is because about 0.0065% of the neutrons resulting from a fission do not get emitted immediately, but instead, are released with a little bit of delay by radioactive fission products that emit neutrons as they quickly decay. So, there is a really abrupt change in the rate of power increase when the fission chain reaction becomes "critical" on just the fraction of the neutrons that are released immediately with the fission of a uranium nucleus. That is called "prompt critical" and is what happened in the Chernobyl reactor accident in Russia.
The "natural reactors" in Oklo, Africa, occurred long ago when there was more U-235 than today, and were apparently self-limiting. For one thing, they were the type of reactor that has the neutrons slowed down by bouncing off hydrogen atoms, with the slower neutrons having a higher probability of causing a fission when they hit a uranium atom than if they were going as fast as they started with the fission event. If there is no water around to slow the fission neutrons before they hit another uranium atom, the concentration of uranium needed to make a "fast (neutron) reactor" critical is much higher.
What the concentration of uranium is in the Earth's core, and how the physics of fission behaves at such immense pressures and temperatures, is somewhat speculative.