Reviews: A Probabilistic Model of Social Decision Making based on Reward Maximization

This is an interesting modeling and model-based analysis study, providing insights into the machinery of human social decision making and possibly its neural correlates. The paper is generally well written and combines advances that could be interesting to both experimental and modeling audience. However, some of its aspects (particularly interpretation of estimated parameters and fMRI analyses) should be improved for it to be acceptable to NIPS. More specific comments: - What is the meaning of groups of 5 individuals if computer generates actions of 19 others? More details about how these actions are generated would be helpful. What are the free parameters/their number for each model?