SHROOM-Visions 2026 asks systems to mark the exact characters hallucinated in a vision–language model’s response. The generating model is withheld, so any model-derived signal is a proxy, and we ask whether it improves character-level localization. We first characterize the metric: span position dominates attainable Cor under the benchmark metric, and constant scores within a response give no ranking signal; an oracle sentence-level detector reaches Cor 0.476 where gold spans reach 0.970. Testing three proxy signals and one direct adjudication baseline, sentence-level signals add at most 0.002 Cor, as predicted. Representation shift is not explained by that account: a probe recovers a fine-grained signal across all four languages (macro AUROC 0.652) yet adds only +0.009. Its discrimination is uneven across error categories, strongest on smaller ones such as OCR errors, but reweighting that mixture accounts for at most 0.028 AUROC. The tested factors do not explain the remaining gap.

We describe the USP system for the AmericasNLP 2026 Shared Task on Culturally Relevant Image Captioning for Indigenous Languages, covering Guaraní (grn), Maya Yucateco (yua), Nahuatl (nah), Wixárika (hch), and Bribri (bzd). We propose a two-stage cascade: Qwen3-VL-8B-Instruct (Bai et al., 2025) generates Spanish captions via language-specific cultural prompts; language-specific fine-tuned NLLB-200-distilled-600M (NLLB Team et al., 2022) models then translate them into each target language. We train on AmericasNLP 2023 data (Ebrahimi et al., 2023) augmented with public parallel corpora. Our system achieves competitive results, including 3rd place in Guaraní human evaluation (2.41/5.0) and 5th in Bribri (1.09/5.0) among 8 teams. We also report that NLLB-200-distilled-600M silently lacks vocabulary entries for Bribri and Maya Yucateco, producing English output without error.

This research investigates the challenges of translating spatial language using open-source LLMs versus traditional NMTs. Focusing on spatial prepositions like ACROSS, INTO, ONTO, and THROUGH, which are particularly challenging for the EN-PT-br pair, the study evaluates translations using BLEU, METEOR, BERTScore, COMET, and TER metrics, along with manual error analysis. The findings reveal that moderate-sized LLMs, such as LLaMa-3-8B and Mixtral-8x7B, achieve accuracy comparable to NMTs like DeepL. However, LLMs frequently exhibit mistranslation errors, including interlanguage/code-switching and anglicisms, while NMTs demonstrate better fluency. Both LLMs and NMTs struggle with spatial-related errors, including syntactic projections and polysemy. The study concludes that significant hurdles remain in accurately translating spatial language, suggesting that future research should focus on enhancing training datasets, refining models, and developing more sophisticated evaluation metrics.