Much effort has gone into detecting stellar CMEs due to their importance for stellar mass and angular momentum loss, and planetary habitability. But what of their stealthier (flare-free) cousins, the filament eruptions (FE)? We explore using the microvariability in filter observations of the He I 1083 nm line to observe FE in stars. Using MLSO CHIP images as our data, we searched for short duration positive gradients in 10830 intensity up to a semi-stable, post eruption level. Working from a solar FE database, we detected 55% of the eruptions with filaments longer than ~150" - mostly the longer, more opaque filaments. A further 42% showed hints of an eruption, but did not pass our detection threshold. Of the events identified, 73% coincided with actual FEs. Application to stellar data should be feasible; tests are planned. (Contributed talk in splinter session.)