Safe transportation is a key use-case of the 5G/LTE Rel.15+ communications, where an end-to-end reliability of 0.99999 is expected for a vehicle-to-vehicle (V2V) transmission distance of 100-200 m. Since communications reliability is related to road-safety, it is crucial to verify the fulfillment of the performance, especially for accident-prone areas such as intersections. We derive closed-form expressions for the V2V transmission reliability near suburban corners and urban intersections over finite interference regions. The analysis is based on plausible street configurations, traffic scenarios, and empirically-supported channel propagation. We show the means by which the performance metric can serve as a preliminary design tool to meet a target reliability. We then apply meta distribution concepts to provide a careful dissection of V2V communications reliability. Contrary to existing work on infinite roads, when we consider finite road segments for practical deployment, fine-grained reliability per realization exhibits bimodal behavior. Either performance for a certain vehicular traffic scenario is very reliable or extremely unreliable, but nowhere in relatively proximity to the average performance. In other words, standard SINR-based average performance metrics are analytically accurate but can be insufficient from a practical viewpoint. Investigating other safety-critical point process networks at the meta distribution-level may reveal similar discrepancies.