Metraipolis › archive › dc_7b8ff6fc28ae
Attention Maps Know When to Think: Training-Free Difficulty-Aware Chain-of-Thought for Reasoning Segmentation — arXiv:2610.41618
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41618
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-02T04:45:23.319Z
- SHA-256 of the archived text
ccf8310f4e883f521b0a248ab09b6372b0cdef057601bc9c6cd3818f82b3c7a0- Stable URL of this page
https://metraipolis.com/documento/dc_7b8ff6fc28ae
How to cite
Attention Maps Know When to Think: Training-Free Difficulty-Aware Chain-of-Thought for Reasoning Segmentation — arXiv:2610.41618 — arXiv (en), https://arxiv.org/abs/2610.41618 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7b8ff6fc28ae (Metraipolis archive dc_7b8ff6fc28ae, sha256 ccf8310f4e883f521b0a248ab09b6372b0cdef057601bc9c6cd3818f82b3c7a0)
BibTeX
@misc{dc7b8ff6fc28ae,
title = {Attention Maps Know When to Think: Training-Free Difficulty-Aware Chain-of-Thought for Reasoning Segmentation — arXiv:2610.41618},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41618},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7b8ff6fc28ae; sha256 ccf8310f4e883f521b0a248ab09b6372b0cdef057601bc9c6cd3818f82b3c7a0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41618. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Chain-of-Thought (CoT) prompting has become a standard reasoning strategy in Multi-modal Large Language Models (MLLMs), showing strong performance in complex vision-language tasks such as reasoning segmentation. However, our analysis reveals that indiscriminate CoT usage can lead to degraded performance by inducing unnecessary reasoning, a phenomenon we term overthinking. In reasoning segmentation, overthinking often leads to misalignment between the query's intent and visual grounding. To address this, we propose Difficulty-Aware Chain-of-Thought (DA-CoT), a training-free adaptive framework that selectively triggers CoT based on the input's intrinsic difficulty. We find that the entropy of the model's attention maps reliably reflects task difficulty: low entropy indicates confident, localized grounding, while high entropy suggests uncertainty and potential benefit from CoT. DA-CoT uses this entropy signal to determine whether to apply CoT prompting. Additionally, we introduce a Bidirectional Rethinking mechanism that enables the model to automatically flip its reasoning decision when the initial output fails, enhancing robustness by recovering from failures induced by overthinking and underthinking. DA-CoT demonstrates its effectiveness by achieving state-of-the-art performance on the Reasoning Segmentation benchmarks while significantly improving inference efficiency. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.41618 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-02T04:42:12.472Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).