How does having a large mempool and allowing a greater transaction ancestry set change the interaction with your peers?

[Bitcoin]

How does having a large mempool and allowing a greater transaction ancestry set change the interaction with your peers?

Bitcoin Core by default allows up to 300 MiB of mempool data, and restricts unconfirmed transaction trees to an ancestry set of at most 25 transactions and 101 kvB of transaction data. Since these are config values, you can obviously use other values than these. How does this change your node's interaction with its peers? Do peers send data that exceeds your node's preferences and your node drops that data upon arrival or does your node inform its peers what to send? Is this the same for mempool data generally and unconfirmed chains specifically? If you allow more data via higher values, does your node forward previously unacceptable data to its peers once their backlog clears enough for the data to be accepted by them?

Source: How does having a large mempool and allowing a greater transaction ancestry set change the interaction with your peers?