A fundamental assumption in paleomagnetism is that a geocentric axial dipole (GAD) geomagnetic field structure extends to the ancient field. Global paleodirectional compilations that span 0 - 10 Myr support a GAD dominated field structure with minor non-GAD contributions, however, the paleointensity data over the same period do not. In a GAD field, higher latitudes should preserve higher intensity, but the current database suggests that intensities are independent of latitude. To determine whether the seemingly “low’ intensities from Antarctica reflect the ancient field, rather than low quality data or inadequate temporal sampling, we have conducted a new study of the paleomagnetic field in Antarctica. This study focuses on the paleomagnetic field structure over the Late Neogene. We combine and re-analyze new and published paleodirectional and paleointensity results from the Erebus volcanic province to recover directions from 107 sites that were both thermally and AF demagnetized and then subjected to a set of strict selection criteria and 28 paleointensity estimates from specimens that underwent the IZZI modified Thellier-Thellier experiment and were also subjected to a strict set of selection criteria. The paleopole (205.6$^{\circ}$, 87.1$^{\circ}$) and $\alpha_{95}$ (5.5$^\circ)$ recovered from our paleodirectional study supports the GAD hypothesis and the scatter of the virtual geomagnetic poles is within the uncertainty of that predicted by TK03 paleosecular variation model. Our time averaged field strength estimate, 33.01 $\mu$T $\pm$ 2.59 $\mu$T, is significantly lower than that expected for a GAD field estimated from the present field, but consistent with the long term average field.