Terrestrial cosmogenic nuclide (TCN) concentrations in fluvial sediment, from which denudation rates are commonly inferred, can be affected by hillslope processes. TCN concentrations in gravel and sand may differ if localized, deep-excavation processes (e.g. landslides, debris flows) affect the contributing catchment, whereas the TCN concentrations of sand and gravel tend to be more similar when diffusional processes like soil creep and sheetwash are dominant. To date, however, no study has systematically compared TCN concentrations in different detrital grain-size fractions with a detailed inventory of hillslope processes from the entire catchment. Here we compare concentrations of the TCN 10Be in 20 detrital sand samples from the Quebrada del Toro (southern Central Andes, Argentina) to a hillslope-process inventory from each contributing catchment. Our comparison reveals a shift from low-slope gullying and scree production in slowly denuding, low-slope areas to steep-slope gullying and landsliding in fast-denuding, steep areas. To investigate whether the nature of hillslope processes (locally excavating or more uniformly denuding) may be reflected in a comparison of the 10Be concentrations of sand and gravel, we define the normalized sand-gravel index (NSGI) as the 10Be-concentration difference between sand and gravel divided by their summed concentrations. We find a positive, linear relationship between the NSGI and median slope, such that our NSGI values broadly reflect the shift in hillslope processes from low-slope gullying and scree production to steep-slope gullying and landsliding. Higher NSGI values characterize regions affected by steep-slope gullying or landsliding. We relate the large scatter in the relationship, which is exhibited particularly in low-slope areas, to reduced hillslope-channel connectivity and associated transient sediment storage within those catchments. While high NSGI values in well-connected catchments are a reliable signal of deep-excavation processes, hillslope excavation processes may not be reliably recorded by NSGI values where sediment experiences transient storage.