The interpretation of experimental results concerning insulator‐semiconductor interface states in metal‐insulator‐semiconductor structures obtained with the deep‐level transient spectroscopy (DLTS) techniques, particularly in the case of high‐density interface states continuously distributed in the energy band gap (larger than 5×1011eV−1 cm−2), is reconsidered. It is shown that the ‘‘saturating pulse’’ condition, which allows a classical treatment of the DLTS spectra, corresponds to a filling pulse amplitude which rapidly increases with the average density of the traps located at the interface. The use of large or small pulses is discussed. The determination of the profile of interface state densityNss(E) can only be derived for high densities from a simulation of the DLTS signal &Dgr;C(T), since the classical relation betweenNssand &Dgr;Cis no more valid in this case. A simplified simulation is proposed. It allows us to justify the results reported in this paper and to fit experimental results previously obtained on Al–Si3N4–GaAs structures with interface state densities about 1013eV−1 cm−2.