The evidence is a little slim (no pun), but there's one theory which I find quite compelling. The theory is that eating foods with high flavor-calorie association increases your set point, and eating foods with low flavor-calorie association decreases your set point.
Flavor-calorie association is when your body knows that a particular flavor results in getting lots of calories in the near future. The flavor-calorie association will be higher for foods that are more flavorful or more calorically dense, and flavor-calorie associations can only be developed for foods you eat regularly. This explains why I lose a bit of weight whenever I go somewhere new – I'm trying new foods, so what I'm eating has doesn't have any flavor-calorie association, so my body's set point goes down, so I get full more quickly, causing me to eat fewer calories overall, causing me to lose weight. And since foods have gotten much more flavorful and calorically dense in the past half-century, it would explain why people's set points have gone up.
There's a simple way to test this theory at home. Extra-light olive oil (not extra-virgin) is tasteless and textureless. So, find a 2-hour window of your day when you don't taste anything with flavor, and have two tablespoons of extra-light olive oil right in the middle of that window. Since the oil has no flavor, it should have no flavor-calorie association, and your set point should go down. I've been doing this, and I've been getting full faster and that's been causing me to lose weight. (If you want to google more info, it's called the shangri-la diet.)
(Note that if you try this, your toothpaste probably counts as a flavor.)