(1) We propose new diagnostics that utilize the [O IV] 25.89 um and nuclear (subarcsecond scale) 12 um luminosity ratio for identifying whether an AGN is deeply “buried” in their surrounding material. Utilizing a sample of 16 absorbed AGNs in the Swift/BAT catalog observed with Suzaku, we find that AGNs with small scattering fractions (< 0.5%) tend to show weaker [O IV]-to-12 um luminosity ratios than the average of Seyfert 2 galaxies. This suggests that this ratio is a good indicator for identifying buried AGNs. Then, we apply this criterion to 23 local ultra/luminous infrared galaxies (U/LIRGs) in various merger stages hosting AGNs. We find that AGNs in most of mid- to late-stage mergers are buried, while those in earlier stage ones are not (Yamada+19). (2) Next, we analyze the broadband X-ray spectra (0.4-70 keV) of the AGNs in two “non-merging” LIRGs (NGC 5135 and UGC 2608) utilizing the data of NuSTAR, Suzaku, XMM-Newton, and Chandra. Applying the X-ray clumpy torus model (XCLUMPY: Tanimoto et al. 2019), we find that both sources show similar spectra characterized by heavily absorption (NH > 1 x 10^24 cm-2) and small torus angular-width (< 30 degrees). We suggest that their tori are not deeply “buried". On the basis of these two results, we argue that the fraction of "buried" AGNs in U/LIRGs increases as the galaxy-galaxy interaction becomes more significant.