Funny and revealing hallucinations
Author: AgatheBauerCreated Aug 4, 2024Updated Aug 21, 2026
While building a GUI, I stumbled across some funny and revealing hallucinations. The GUI contains a microphone mode. While saying nothing or something completely different, I'm often getting transcription snippets like this one:
"Untertitel im Auftrag des ZDF, 2020"
ZDF is a German television channel. Looks like the model was trained with data from it... :D
Is there a way to fix this?
If anyone is interested, I attached the code. Completely made with GPT4o: Main.zip
Source: SYSTRAN/faster-whisper