The average trained cyclist has an FTP of roughly 80% of VO2max. You can therefore approximate the percentage of FTP by dividing by 0.8 (or by multiplying by 1.25).
For example, if you read a classic study where participants exercised at 70-75% of VO2max for 4 h when fed carbohydrate at >100 g/h 
, that would have been approximately 88-94% of FTP. Similarly, Seiler’s “lower pole” around which training in his zone 1 (of 3) of roughly 60% of VO2max would be about 75% of FTP.
Less than about 70% of FTP is unlikely to induce further adaptations in someone who is already training regularly (unless perhaps carried out for many, many hours). As the intensity is progressively increased, so too is the stimulus for adaptation. There is therefore no downside to going harder (e.g., at level 3), except that it is more fatiguing. The latter is why when most people do an “endurance ride” (“LSD”, whatever) outdoors, they end up with an average power of 56-75% of FTP.* Of course, the longer the workout, the more likely it will be towards the bottom end of that range, and the shorter the workout, the more likely it will be towards the top of that range, even perhaps slightly above. For example, I did my 1 h (really 50 min once you subtract warmup and cool down) “moderate intensity filler workouts” at 83% of FTP. I never went easier than that indoors, except on the rare occasion that I felt like I needed a recovery ride, and/or I had a race in the next couple of days (then it was 50% of FTP for 1 h…never longer!).
*As discussed previously, when cycling outdoors, power tends to be highly variable, which pulls down the average. It’s really better to think in terms of IF in this context, i.e., by envisioning the effects of an isopower workout relative to FTP.
TL,DR: It’s your glycogen budget - spend it wisely.