I have a pyrochlore antiferromagnetic system. In AC magnetic susceptibility measurements, there is a cusp at 4 K whose frequency dependence follows a spin-glass behavior. There is a cusp at around 5 K in magnetic heat capacity, also suggesting a spin-glass transition. But in neutron scattering, I observed magnetic correlations and dynamic spin fluctuations at 50 mK (well below 4 K), which maynot be reasonable for a spin glass with spin-glass temperature around 4 K. Some poeple told me this system should be more spin-liquid-like, rather than spin-glass. I am wondering whether there is a distinct experimental fact that can separate spin glass and spin liquid?