Seems like you could test that via negative stimulus. Introduce a negative stimulus like a shock associated with something perceivable by both the caterpillar and butterfly. After conditioning the caterpillar, see if the butterfly shares the same aversion to the stimulus.

This has been tested! In moths at least:

> Manduca sexta caterpillars received an electrical shock associatively paired with a specific odor in order to create a conditioned odor aversion, and were assayed for learning in a Y choice apparatus as larvae and again as adult moths. We show that larvae learned to avoid the training odor, and that this aversion was still present in the adults.

https://journals.plos.org/plosone/article?id=10.1371/journal...

tldr summary; moth doggos retain learnings from pupper form, owo :3

The article mentions that memories are maintained - I didn’t see a source, but here is an article about it: https://www.newscientist.com/article/dn13412-butterflies-rem...

The caterpillar turns to mush. Its nervous system seamlessly morphs into a new one. Many neurons are retained, some are pruned, and some new ones are formed.

This is why certain memories can be retained and certain ones can't.

I remember being taught this as a child and iirc recent studies say it doesn’t mush.

EDIT:

"Textbooks will commonly talk about the insect dissolving into a kind of “soup”, but that’s not entirely accurate. Some organs stay intact. Others, like muscles, break down into clumps of cells that can be re-used, like a Lego sculpture decomposing into bricks. And some cells create imaginal discs—structures that produce adult body parts. There’s a pair for the antennae, a pair for the eyes, one for each leg and wing, and so on. So if the pupa contains a soup, it’s an organised broth full of chunky bits."

https://www.nationalgeographic.com/science/article/3-d-scans...