Table 1. Model parameter settings.

These were held constant across all simulations and experiments.

Parameter	Range	Role	Value
α	[0, + ∞)	CRP concentration parameter: larger values favor more clusters	1
ϵ	[0, 1]	choice stochasticity: larger values lead to more deterministic choices	0.6
# samples	[1, + ∞)	number of MCMC iterations per simulated participant	10000
$\overline{\theta }$	(−∞, + ∞)	average reward of entire graph	15
σθ	(−∞, + ∞)	standard deviation of average cluster rewards around $\overline{\theta }$	10
σμ	(−∞, + ∞)	standard deviation of average state rewards	10
σr	(−∞, + ∞)	standard deviation of state rewards	5