Density Checker Hates Content Optimization?

okay, so our 'Keyword Density & Frequency Checker' web tool is being a bit dramatic lately. and it's not a good one. it's designed for content optimization, showing keyword distribution for SEO, but it's gotten seriously confused.

i'm talking about basic stuff here. you feed it a normal piece of text, something with common english words, and it's like "nah, 'the' has 0% density." zero! but then some random, obscure word that appears maybe once gets like, 200%. how does that even happen? i've checked the parsing logic, the tokenization, everything seems normalish? but the output is clearly not. it's like it's counting words that aren't there or multiplying the count of others.

here's a little peek at what i'm seeing in the console sometimes, just to illustrate the madness:

--- Keyword Analysis Report ---
Word: "the" | Count: 0 | Density: 0.00%
Word: "and" | Count: 0 | Density: 0.00%
Word: "flibbertigibbet" | Count: 2 | Density: 215.38%
Word: "serendipity" | Count: 1 | Density: 107.69%

has anyone else had their text analysis tools go rogue like this? i'm really scratching my head for debugging tips or if there's some common pitfall in freq/density calcs that i'm totally missing. maybe it's a weird character encoding issue i'm overlooking? any insights would be a life saver. thanks in advance!