Whisper’s 31% Silent-Input Hallucinations Drop Near 2% Without Retraining
Whisper can confidently transcribe speech that was never there, with hallucinations hitting 31.31% on non-speech inputs. This training-free method learns a compact hallucination-associated subspace from silent calibration data, then projects it out of the decoder at inference time, cutting the rate to 2.44%. A gated version reaches 3.74% while trading a 0.33 to 4.39 point WER increase on genuine speech, giving production systems a tunable safety-performance knob.