Non-characters get mistaken as UTF-8 characters

[sati-bodhi]

I have the following Chinese word list as part of my basic.json array. [ "一一", "一一〇", "一一九", "一丁", "一丁不識", "一丁點", "一丈差九尺", "一丈青", "一三五不論，二四六分明", "一上", "一下", "一下子", "一不做，二不休" ] When typing non-character symbols like -``=``\, auto-complete gets triggered even when fuzzy input has been disabled on the auto-complete-plus package.

Is there a way to work out of this by adding code or something?