From 4cb8f2fd53945ddc5c86a9e09474516eb3a249f6 Mon Sep 17 00:00:00 2001 From: "jinli.yl" Date: Fri, 16 Oct 2026 11:12:26 +0800 Subject: [PATCH] docs: fix compaction benchmark links with verified permalinks --- docs/en/context_compaction.md | 2 +- docs/zh/context_compaction.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/en/context_compaction.md b/docs/en/context_compaction.md index daf04be8..3e534d57 100644 --- a/docs/en/context_compaction.md +++ b/docs/en/context_compaction.md @@ -120,7 +120,7 @@ latency and total cost alongside token reduction. ReMe does not automatically su ## Span-selection experiment -The [benchmark protocol and results (Chinese)](https://github.com/agentscope-ai/ReMe/blob/main/benchmark/compaction/RESULTS.md) +The [benchmark protocol and results (Chinese)](https://github.com/jinliyl/ReMe/blob/cf5587681070735db6b49fe90f93c752a17935eb/benchmark/compaction/RESULTS.md) describe a local run on October 15, 2026, using a synthetic Hermes payment trajectory and a controlled long-log augmentation. It compares full evidence, Jev span selection, head/tail clipping and ordinary Qwen summaries under 600/1000 evidence-token allowances, with identical system/user instructions and evaluation prompts. diff --git a/docs/zh/context_compaction.md b/docs/zh/context_compaction.md index a1406b4c..b6be29c5 100644 --- a/docs/zh/context_compaction.md +++ b/docs/zh/context_compaction.md @@ -111,7 +111,7 @@ ReMe 不会自动生成摘要或写入长期记忆。 ## 分块压缩实验 -[Benchmark 评测方案与结果](https://github.com/agentscope-ai/ReMe/blob/main/benchmark/compaction/RESULTS.md) +[Benchmark 评测方案与结果](https://github.com/jinliyl/ReMe/blob/cf5587681070735db6b49fe90f93c752a17935eb/benchmark/compaction/RESULTS.md) 记录了 2026-10-15 的本地对照实验：以 Hermes 采集的合成支付 trajectory 及其长日志增噪变体，比较完整证据、 Jev 分块筛选、首尾截取和普通 Qwen 摘要。证据预算为 600 / 1000 token，所有组保持相同的 system、user 和评测提示。 分块排序只在实验脚本中实现；上述 `compact_context` 的配对判断与低概率预览策略保持不变。