Injecting self-verified bounding boxes into Chain-of-Thought data improves few-shot adaptation of multimodal LLMs on charts, tables, receipts, and reports.
Lora: Low-rank adaptation of large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Bootstrapping Grounded Chain-of-Thought in Multimodal LLMs for Data-Efficient Model Adaptation
Injecting self-verified bounding boxes into Chain-of-Thought data improves few-shot adaptation of multimodal LLMs on charts, tables, receipts, and reports.