无需训练，用图结构压缩长文本，效果好且省资源。
TopoCompress: Long Context Compression via Graph-Wired Semantic Trajectories
- 通过语义轨迹图选择连贯语义片段，不依赖特定模型。
- 在5个任务中表现优于基线，压缩量减少4倍仍保持高精度。
- 适合追求高效推理的LLM部署场景，尤其适合资源受限环境。
长上下文压缩对降低大语言模型推理成本和延迟至关重要。然而，现有方法常导致关键证据断裂、需额外训练或对目标模型依赖性强。我们提出TopoCompress，一种无需训练、模型无关的框架，通过选择连贯语义段来压缩长上下文。该方法首先结合密集查询相关性与词法查询相关性，以及语义加速，为每个段落打分；随后构建混合图，基于语义相似性和序列邻近性连接段落，并在图上传播查询引导的相关性分数。在五个长上下文任务（HotpotQA、2WikiMQA、MuSiQue、Qasper、MultiFieldQA）上，TopoCompress始终优于强基线。值得注意的是，其性能接近最强基线，同时压缩预算仅为后者的1/4，且压缩时间比最快基线减少1.41倍。
原文摘要 · Abstract (English)
Long-context compression is essential for reducing the cost and latency of large language model inference. However, existing methods can fragment important evidence, require additional training or alignment, and often depend on the target model for effective compression. We introduce TopoCompress, a training-free and model-agnostic framework that compresses long contexts by selecting coherent semantic spans. TopoCompress first scores each span using dense and lexical query relevance together with semantic acceleration. It then constructs a hybrid graph that connects spans based on semantic similarity and sequential adjacency, and propagates the query-guided relevance scores over the graph. Across five long-context tasks-HotpotQA, 2WikiMQA, MuSiQue, Qasper, and MultiFieldQA-en-TopoCompress consistently outperforms strong compression baselines. Notably, TopoCompress achieves performance comparable to the strongest baseline while using a 4x smaller compression budget, and provides a 1.41x smaller compression time over the fastest baseline.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。