There are several open issues related to Gaussian splatting. It could be helpful to add some screenshot here, to narrow down the possible reasons: Are the splats looking “fuzzy”, or are there inexplicable “lines/bands” of splats?
Based on the description until now (with the broad description of being from a PLY and being displayed “incorrectly”), the issue of Regression in splats from PLY in 1.135 · Issue #13041 · CesiumGS/cesium · GitHub could be hot candidate. If this is it, then I recently posted a workaround for that in Gaussian Splats Apparently Breaking - #2 by Marco13 . Otherwise, it may be necessary to investigate this further.