Utilitarian ethical theories state that the morality of actions depends on their consequences. Recent work suggests that individual differences in utilitarian tendencies fall along two dimensions: a permissive attitude towards harming others for some greater good (instrumental harm) and an impartial concern for others’ welfare (impartial beneficence). Here, we investigated whether these two dimensions are dissociable in behavior and brain activity during moral decision-making. In studies of both incentivized and hypothetical decisions to earn money by inflicting pain on oneself or a stranger (N=79 and N=192), instrumental harm positively predicted harming others, while impartial beneficence negatively predicted harming others relative to oneself. Inter-subject representational similarity analysis of functional neuroimaging data (N = 68) revealed that participants with similar utilitarian tendencies showed similar neural encoding of moral choice attributes, even if they made different choices. These similarity patterns emerged in distinct brain regions for instrumental harm and impartial beneficence. Our results provide behavioral and neural support for a two-dimensional model of utilitarianism that dissociates instrumental harm from impartial beneficence and suggest that a commitment to particular moral principles predicts neural representations of moral choice attributes over and above the actual decisions people make.