It's possible to see how many factors there are in a result file; but it's not possible to tell how many of those are actually relevant. See, since the PPSE sieve file is so big, it's not sent out with the work. Instead, your computer returns all the factors it finds in the given range, excluding those that factor a number already divisible by a small prime, and the rest get sorted out by the validator.
I've checked, and it would be too much work to sort out how many relevant factors there were in each WU. However, it would be interesting to get an aggregate average of what percentage of factors are relevant now. At the 40T level, it was about 1 in 4 returned factors; but I don't know what it is now.
John? Rytis? Can you provide some numbers for average relevant factors per WU?