Which configuration file will be used when run CI on merge

Hi, in our repository (hosted on private gitlab) we have master branch and dev branch. The configuration file is written by jsonnet looks like this:

local validate() {
  kind: 'pipeline',
  type: 'kubernetes',
  name: 'validate-pull-request',
  steps: [
  trigger: {
    branch: 'master',
    event: 'pull_request'
validate (),

The we submitted a pull request to merge dev to master. In the dev branch, the CI configuration file is an older version which has less check rules than HEAD ver in the master.

We found that the CI didn’t use the merged configuration file to run the CI, but just uses the configuration in the dev branch even which is older than master.

May I ask help to resolve this problem?

Thank you!