Two-Token Features and Small-Large Ensembles for VLM Hallucination Detection

We present our system for the SHROOM-Visions 2026 shared task on character-level VLM hallucination detection. A small (4 B-parameter) VLM is fine-tuned as a per-token classifier reading a two-token feature from its own hidden states, and is ensembled with a about 400B zero-shot VLM judge at prediction time. Both components see off-the-shelf OCR of any visible in-image text. We use synthetic hallucination data generated by the large model as a source of ensemble diversity, and use validation to select feature layer, training data and OCR grounding. Our official entry reaches mean Cor 0.487 / Cor-lbl 0.387 on the hidden test set, placing 6 th/ 28 (EN), 6 th/ 21 (FR), 8 th/ 21 (IT) and 7 th/ 22 (ZH) on the task's primary Cor-lbl metric.