Stochastic analyses are conducted of model neural networks of the generalized Little-Hopfield Hemmen type, in which the synaptic connections with linearly embedded p sets of patterns are free of symmetric ones, and a Glauber dynamics of a Markovian type is assumed.