Qiyan Zhao
Qiyan Zhao has 2 papers in our library, in 2026; often with Xiaofeng Zhang, Jiawei Cao, Ruixuan Xu.
OpenAlex · h-index 0 · 2 works · 0 citations OpenAlex, 2026-10-14
Affiliations
Fields, topics and concepts
📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 2 worksComputer Vision and Pattern Recognition · 2 worksArtificial Intelligence · 1 workArts and Humanities · 1 workLanguage and Linguistics · 1 work
Works by year
2020s
Papers
- Fixing Semantic Blind Spots in Anchor Tokens of dMLLMs (2026)
Diffusion-based multimodal LLMs (dMLLMs) such as LLaDA-V and LaViDa still describe things that are not in the image. - TokenPenalty: Alleviating Attention Sinks and Positional Decay in LVLMs (2026)
Large vision-language models (LVLMs) often mention objects that are not in the image. This undermines trust in uses such as captioning and product descriptions. The authors link part of this to two effects. First, attention collapses onto a few visual "sink" tokens with extreme query/key activations. Second, RoPE…
Co-authors here
Venues
- Findings of the Association for Computational Linguistics ×2: Findings of the Association for Computational Linguistics 2026 ×2
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/authors/qiyan-zhao/