I am wondering if my participants' responses will still be correlated to token IDs if I use dummy tokens to ensure pseudonymised participation?

Dummy tokens are just a way to create a participants list without entering a lot of data.
The token functionality of a dummy token is in no way different than a token created manually with an email and personal data.

For every token usage:
If you run an anonymized survey, the token data is not saved in the response set.

Thank you! So for a pseudonymised survey, the token data IS saved in the response set?

Every token (no matter how you have generated them) is saved into the responseset (as long the survey is NOT set to anonymized).