Though the existence of two-level systems (TLS) is widely accepted to explain low temperature anomalies in many physical observables, knowledge about their properties is very rare. For silica which is one of the prototype glass-forming systems we elucidate the properties of the TLS via computer simulations by applying a systematic search algorithm. We get specific information in the configuration space, i.e. about relevant energy scales, the absolute number of TLS and electric dipole moments. Furthermore important insight about the real-space realization of the TLS can be obtained. Comparison with experimental observations is included.PACS numbers: 61.43.Fs, 63.50.+x, Most kinds of disordered solids show anomalous behavior at very low temperatures (Kelvin regime and below) as compared to their crystalline counterparts. Many of the observed features can be explained by the Standard Tunneling Model (STM) [1,2] and its generalization, which is the Soft-Potential Model [3,4]. The basic idea of the STM is to postulate the possibility of localized transitions between different configurations, i.e. adjacent minima of the potential energy landscape. Such a transition can be described by a double-well potential (DWP), characterized by an asymmetry ∆, potential height V and distance d between both configurations. From a quantum-mechanically perspective at low temperatures the system is tunneling between both configurations and the DWP is characterized by the lowest two eigenstates. If their energy difference E is in the Kelvin regime, these DWP may contribute to the low-temperature anomalies. Then one may speak of Two Level Systems (TLS). The TLS can couple to strain and electric fields and therefore show up in observables like thermal conductivity, sound absorption and dielectric response [5]. Recently, even observations about the interaction of different TLS have been reported [6,7,8].So far it has not been possible to derive a theory of the low-temperature anomalies of real glass-forming systems from first principles, except for mean-field models [9] and random first order transition theory [10]. Thus for a prototype system like SiO 2 the STM has to be basically considered as phenomenological. Important questions emerge. Computer simulations may help to shed some light on the nature of the low-temperature anomalies. In previous work on SiO 2 the trajectories, generated either by molecular dynamics [11,12] or by the activationrelaxation technique [13], have been analyzed with respect to transition events between different structures. Both approaches yield some interesting insight into the nature of relaxation processes in SiO 2 . (Q1), however, requires a systematic search procedure, and (Q2) and (Q3) a sufficiently large number of characteristic DWP and thus an efficient search method. This has not been the scope of previous simulations on SiO 2 .In recent years we have developed a set of simulation techniques which allow us to approach these questions [14,15]. (Q1) Starting from representative low-energy structures we...