Tautologies can only be true within a given logical framework,
Agreed. I'm not sure that changes anything.
Experimental facts, or empirical knowledge, can also never be said to be even provisionally true
It sounds like you're referring to Hume's problem of induction.
5 sigma is a measure of how confident scientists feel their results are. If experiments show results to a 5 sigma confidence level, that means if the results were due to chance and the experiment was repeated 3.5 million times then it would be expected to see the strength of conclusion in the result no more than once.
[LINK]
Most published "studies" never reach 2 sigma.
There are very well established standards of evidence, although most people are not aware of them and most claims overstate their own confidence.
Give me any true empirical fact.
The planet earth exists.