As crash avoidance technologies like autonomous emergency braking (AEB) become more prevalent, it is increasingly necessary to study occupant responses during these scenarios to understand their implications on occupant safety. The low speed and long duration of pre-crash braking and low severity crash events provide sufficient time for muscle activation to affect occupant response. Unfortunately, this effect is not accurately captured in post mortem human subject response data from high severity tests that are typically used to validate anthropomorphic test devices (ATDs). Performing matched low-speed volunteer and ATD tests can provide insight into the validity of using ATDs to assess occupant safety during low severity events. The THORAV-5F ATD was developed recently and was designed to represent 5th percentile females in autonomous vehicles. Its biofidelity has not yet been assessed in low severity tests. The objective of this study was to quantify the occupant response of the THOR-AV-5F during low-speed frontal and frontal-oblique sled tests, and compare the ATD response to relaxed and braced 5th percentile female volunteer responses during matched tests. Six 5th percentile female volunteers experienced low-speed sled tests in two test orientations (frontal and frontal-oblique) and at two pulse severities (1 g and 2.5 g), with two muscle conditions (relaxed and braced) per pulse severity. Matched THOR-AV-5F sled tests were performed for each test condition. Reaction forces at each test buck interface and head, neck, and sternum accelerations were measured for the volunteers and ATD during each sled test. Overall, the THOR-AV-5F reaction forces and accelerations exhibited some differences compared to the volunteers, but were generally more similar to the relaxed volunteers than to the braced volunteers across all test conditions.