Steps Towards the Pluralistic Alignment of Language Models