Jul 2026
The Count Is There, but Misaligned: Understanding and Correcting Counting Failures in VLMs
A detector-guided self-correction method is proposed that selectively re-prompts the model only when an internal error detector predicts failure, and a simple inference-time intervention improves counting accuracy by up to 15.6 absolute percentage points.
Ahmed Oumar El-Shangiti, Abzal Nurgazy, Hilal AlQuabeh et al.
· arXiv.org · 1 citation