Are there scientific theories, maybe string theory or the multiverse, where you would not call the empirical data they use, verified by scientific methods, “evidence presented in support of those theories” simply because those theories are not testable? Since that is the only word you use that distinguishes #2 from #1.
I know some scientists do consider string theory or the multiverse to be non-testable pseudoscience or philosophy for these same reasons. But just want to see if there is consistency in your view. If so, that’s fine.
If the the empirical data, verified by scientific methods, is not disputed and presented as evidence for a claim, it doesn’t seem like those who don’t think the evidence is strong or that dispute the claim, should be able to dismiss it as not even being called “evidence” in support of a claim. Sure, they can say it doesn’t support the claim.
The data collectively known as “fine tuning” (though I realize some don’t like that term) seems to be a well-established accumulation of empirical data verified by scientific methods. Some may use it as evidence for a multiverse, some an intelligent creator. We may not agree with their claims, but seems valid to allow it to be called “evidence” by the claimants in either case (or no case, if that is your view).