Solar filaments are prominent cold and dense structures of plasma suspended in the hot corona. They are called prominences when observed above the solar limb. Despite they were probably the first solar features to be observed, the magnetic structure of solar filaments is still largely unknown, being their determination a fundamental problem for Solar Physics. The reason is the lack of chromospheric observables able to provide binding information for the magnetic field. Therefore, though there are two main proposed models, namely sheared-arcade and twisted flux-rope models, due to the lack of observational constraints it is unclear what is the filament most probable magnetic configuration. High-spatial resolution observations have revealed that filaments are also highly structured at very small scales. Remarkably, most solar filaments erupt, giving rise to coronal mass ejection. Thus, the understanding of solar filaments is also important for space-weather forecasting. To infer their magnetic structure it is crucial to observe the formation and evolution of solar filaments in the Photosphere and in the Chromosphere, in polarized light, and at the highest spatial resolution. We propose to observe solar filaments with the Tenerife Infrared Polarimeter (TIP-II) installed at the GREGOR ground-based 1.5-meter telescope in synchrony with the Hinode space-based observatory. I will use TIP-II to observe the 1083.0 nm spectral region containing a photospheric line and the chromospheric He I 1083.0 nm triplet. This line is sensitive to atomic level polarization and to the joint action of the Hanle and Zeeman effects. Thus, it is ideal for determining a wide range of field strengths in prominences. The data will be analyzed using state-of-the art inversion algorithms. Finally, Hinode will provide a wealth of binding information for the determination of the magnetic structure of solar filaments and also for identify possible trigger mechanism of filament eruption.