Step-Aware Decoding:一种缓解多模态大模型幻觉的新方法
Rethinking in Spikes: Mitigating Hallucinations in MDLMs with Step-Aware Decoding
Step-Aware Decoding 通过在解码步骤中引入上下文感知机制,有效降低了多模态大模型(MDLMs)的幻觉问题。该方法在 MMLU、HumanEval 等基准测试中,幻觉率降低了 15%,同时推理速度提升了 2.1 倍。该技术已开源,支持上下文长度 128k tokens,可通过 API 调用。
Published on Oct 14
Authors:
,
,
,
,
,
,
,
,
,
,
View arXiv page View PDF GitHub 3 Add to collection
Models citing this paper 0
No model linking this paper
Cite arxiv.org/abs/2611.12901 in a model README.md to link it from this page.
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2611.12901 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2611.12901 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
来源: Hugging Face Daily Papers · huggingface.co