Yeah, I see it.
However, actually this does not mean the inputs are passed as environment variables.
By default, GitHub will try to map all the provided inputs of an action as environment variables on the runner (other CI/CD services, such as Azure Pipelines, also are similar).
When run the Docker container actions, it also will try to map the related environment variables from the the runner into the container. But for some reasons, the inputs of the Docker container actions actually are not mapped as environment variables on the runner. So the nonexistent environment variables will not be mapped into the container, although they are set in the docker run command.
I guess, the environment variables for the inputs is designed to store the information (<key>:<value>) of the inputs, so that the scripts executed in the action can use these information to do some things (for example, using in the if statement).
Because the inputs of the Docker container actions have not been mapped as environment variables, currently the environment variables for the inputs you see in the docker run command are not sense. These names of environment variables may currently only be set as backup so that can be used in the future.
If you execute the command “printenv” in the Docker container action, you can see these environment variables actually is not existing in the container (here is an example).
Hope my above explanations can help you understand.