Code editing requires a model to decide where to make changes, generate the new content, and preserve everything else. We study how masked diffusion language models divide these responsibilities across four editing interfaces: whole-file rewriting, search-and-replace, locate-then-infill, and token-level editing. Experi...
Xi-Jia Tao, Zi-Rui Liu, Shansan Gong et al.· 0 citations
Findings show that a skill document can serve as the starting point for a continuous control that improves both answers and actions without updating the backbone.
Xi-Jia Tao, Yi-Hua Teng, Xin-Yu Fu et al.· arXiv.org· 3 citations
VizRAG is introduced, the first RAG system to support visual hypergraph structure awareness, and demonstrates that VizRAG significantly outperforms strong baselines, validating the promising potential of hypergraph visualization as a novel approach for RAG systems.
Yan-Bin Wei, Yang Chen, Ren-Ling Gan et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.