Consider the ideal Carnot cycle consisting of two ideal reversible isothermal stages at $T_0$ and $T_1$ and two ideal adiabatic reversible (isentropic) stages connecting them; assume that $T_0 > T_1$. Denote the corresponding isothermally absorbed and rejected heat transfers, resp., by $Q_0$ and $Q_1$ so the work delivered by the cycle is $W=Q_0-Q_1$.
During the higher temperature isothermal stage the engine receives $Q_0$ thermal energy at temperature $T_0$ and at the same time also absorbs $S_0=\frac{Q_0}{T_0}$ entropy. Similarly, at the lower temperature $T_1$ it rejects $Q_1$ thermal energy and $S_1=\frac{Q_1}{T_1} entropy$. Since there is no entropy change during the two adiabatic reversible (isentropic) stages, not just globally but also locally, we must have $S_1=S_0$.
Now it is usually claimed with Clausius that the incoming heat is simultaneously converted to work. Since the delivered work is $W=Q_0-Q_1$ and at the lower temperature $T_1$ isothermal stage energy is expelled, the conversion is to occur at the high temperature $T_0$ isothermal stage. By this interpretation, the "heat" as flows in is being converted to work and one can only have this by assuming that the total absorbed entropy, $S_0$, inside the engine, is being reduced by $S_x=\frac{W}{T_0}=\frac{Q_0-Q_1}{T_0}$ as the heat to work conversion is taking place.
In other words, by the time the adiabatic stage starts only $S_y=S_0-S_x=\frac{Q_0}{T_0}-\frac{Q_0-Q_1}{T_0}=\frac{Q_1}{T_0}$ excess entropy left relative to what it was at the beginning of the cycle. This does not change in the isentropic stage but it is further reduced by the expelled entropy in the isothermal stage at $T_1$ after which inside the engine there remains $\Delta S= \frac{Q_1}{T_0}-\frac{Q_1}{T_1}$. But this $\Delta S < 0$ because $T_1<T_0$, and since this stage is followed by an isentropic stage to restore the cycle there must be some positive entropy source to compensate for this "missing" entropy; this is the uncompensated heat of Clausius.
In this cycle then there must be an excess production of $-\Delta S$ entropy during the isothermal stage at $T_1$ during which the entropy $S_1=S_0$ along with heat $Q_1$ are expelled. So here the isothermal stage while it is supposedly reversible also generates $-\Delta S >0$ entropy as compensation in addition to transferring reversibly $S_0$ entropy from the engine to the lower temperature reservoir.
Am I the only one here who finds this explanation strange? So, how does compensation really work?