We introduce Stingy Context, a hierarchical tree-based compression scheme achieving 18:1 reduction in LLM context for auto-coding tasks. Using our TREEFRAG exploit decomposition, we reduce a real source code base of 239k tokens to 11k tokens while preserving task fidelity. Empirical results across 12 Frontier models show 94 to 97% success on 40 real-world issues at low cost, outperforming flat methods and mitigating lost-in-the-middle effects.
翻译:本文提出“吝啬上下文”,一种基于分层树结构的压缩方案,在自动编码任务中实现了18:1的LLM上下文压缩比。通过我们提出的TREEFRAG分解方法,一个包含23.9万标记的真实源代码库被压缩至1.1万标记,同时保持了任务完整性。在12个前沿模型上的实验结果表明,该方法以较低成本在40个现实问题中取得了94%至97%的成功率,其性能优于扁平压缩方法,并有效缓解了“迷失在中间”效应。