The well-documented propensity of negatively-chirped pulses to enhance resonant impulsive Raman scattering has been rationalized in terms of a one pulse pump-dump sequence which "follows" the evolution of the excited molecules and dumps them back at highly displaced configurations. The aim of this study was to extend the understanding of this effect to molecules with many displaced vibrational modes in the presence of condensed surroundings. In particular, to define an optimally chirped pulse, to investigate what exactly it "follows" and to discover how this depends on the molecule under study. To this end, linear chirp effects on vibrational coherences in poly-atomics are investigated experimentally and theoretically. Chirped pump-impulsive probe experiments are reported for Sulforhodamine-B ("Kiton Red"), Betaine-30 and Oxazine-1 in ethanol solutions with <10 fs resolution. Numerical simulations, including numerous displaced modes and electronic dephasing, are conducted to reproduce experimental results. Through semi-quantitative reproduction of experimental results in all three systems we show that the effect of group velocity dispersion (GVD) on the buildup of ground state wave-packets depends on the pulse spectrum, on the displacements of vibrational modes upon excitation, on the detuning of the excitation pulses from resonance, and on electronic dephasing rates. Akin to scenarios described for frequency-domain resonance Raman, within the small-displacement regime each mode responds to excitation chirp independently and the optimal GVD is mode-specific. Highly-displaced modes entangle the dynamics of excitation in different modes, requiring a multi-dimensional description of the response. Rapid photochemistry and ultrafast electronic dephasing narrow the window of opportunity for coherent manipulations, leading to a reduced and similar optimal chirp for different modes. Finally, non-intuitive coherent aspects of chirp "following" are predicted in the small-displacement and slow-dephasing regime, which remain to be observed in experiment.