Polish dataset documentation and link the evaluation guide

README.md

@@ -13,40 +13,93 @@ configs:
         path: preview/tasks.jsonl
 ---
 
-# ReScopeBench
+<div align="center">
 
-ReScopeBench evaluates experience formation, reuse, and revision in **18 stateful
-environments, 352 tasks, and 26 environment changes** per setting.
+<h1>ReScopeBench</h1>
+<h3>Towards Robust Long-Lived Agents</h3>
 
-[Homepage](https://rescopebench.github.io/)
+<p><i>Can agents keep experience useful when their environment changes?</i></p>
 
-- `cases/`: executable environments, initial states, task streams, and evaluator assets.
-- `preview/tasks.jsonl`: unchanged public task prompts for browsing, not
-  standalone evaluation inputs.
-- `cases-manifest.json`: the file inventory and SHA-256 checksums.
+[Homepage](https://rescopebench.github.io/) |
+[Code](https://github.com/JoeYing1019/ReScopeBench) |
+[Quick Start](https://github.com/JoeYing1019/ReScopeBench#quick-start) |
+[Main Results](https://github.com/JoeYing1019/ReScopeBench#main-results)
 
-## Use
+<p>
+  <b>18</b> stateful environments &nbsp;&middot;&nbsp;
+  <b>352</b> tasks &nbsp;&middot;&nbsp;
+  <b>26</b> drift events
+</p>
 
-Use the code repository whose `config/cases.json` pins this dataset revision.
-From that repository's root, run:
+</div>
+
+## Overview
+
+**ReScopeBench evaluates the lifecycle of agent experience:** acquiring local
+knowledge, reusing it for new goals, and selectively revising it when the
+environment changes. Its task streams share structural and rule knowledge,
+while unannounced drift events make only part of that knowledge obsolete.
+
+- **Structural experience:** how resources are located, connected, and accessed.
+- **Rule experience:** the local conditions and consequences that govern actions.
+- **Evolving tasks:** streams of 16, 24, or 32 tasks with persistent business
+  state and one, two, or three drift events.
+
+This dataset contains the executable **benchmark cases**. The
+[GitHub repository](https://github.com/JoeYing1019/ReScopeBench) provides the agent
+harness, model API adapters, metric computation, and report generation.
+
+## Dataset Contents
+
+- **`cases/`** contains environment implementations, initial states, task streams,
+  drift definitions, and evaluator assets. Each environment has one `driver.py`
+  and one `formal-binding.json`.
+- **`preview/tasks.jsonl`** exposes the 352 public task prompts for browsing.
+  The `case_path` field identifies the corresponding environment directory.
+- **`cases-manifest.json`** records the file inventory and SHA-256 checksums used
+  by the installer.
+
+The preview is a view of the task streams, **not an independent-question
+evaluation set**. Evaluating a row in isolation does not reproduce the benchmark.
+
+## Run the Benchmark
+
+Follow the [installation guide](https://github.com/JoeYing1019/ReScopeBench#installation)
+to set up Python 3.10, Docker, and the evaluation repository. From its root,
+download the matching cases:
 
 ```bash
 python scripts/download_cases.py
 ```
 
-Follow the code README for dependencies, Docker setup, model configuration, and
-metrics.
+The installer selects the immutable dataset revision pinned in the code
+repository's `config/cases.json` and verifies the downloaded files. Use that
+matched code/data pair rather than combining arbitrary revisions.
+
+For every model and environment, run **both complete settings**:
+
+- **w exp:** retain naturally produced memory, skills, and permitted history.
+- **w/o exp:** clear explicit experience between tasks.
+
+The two streams start from the same initial environment and then maintain their
+own business state. Clearing experience does not reset that state. See the
+[quick start](https://github.com/JoeYing1019/ReScopeBench#quick-start) for model
+configuration, batched tool calls, and execution commands.
+
+Reference knowledge and grading assets belong to the evaluator, not to learner
+inputs. These cases use the ReScopeBench harness; they are not native Harbor
+task bundles. Preserve applicable third-party notices and licenses.
 
-Run each environment as two independent sequential streams: **w exp** retains
-cross-task experience; **w/o exp** clears it at task boundaries. Do not run preview
-rows as independent questions or expose the full stream to the learner.
+## Evaluate Results
 
-Each case contains a `driver.py` and `formal-binding.json`. The preview's
-`case_path` points to its environment directory.
+The [metric pipeline](https://github.com/JoeYing1019/ReScopeBench#metrics) reports
+task performance, structural and rule experience coverage, immediate updating,
+and recovery timeliness, with CSV and HTML outputs. Experience scores measure
+coverage of selected reference knowledge and accept equivalent expressions;
+task graders also allow other valid strategies.
 
-These packages require the matching ReScopeBench framework; they are not native
-Harbor task bundles. Keep reference knowledge, scoring code, and evaluator-only
-state outside learner inputs. Preserve applicable third-party notices and licenses.
+The [main results](https://github.com/JoeYing1019/ReScopeBench#main-results)
+compare 20 models across all 18 environments under both settings.
 
 ## Acknowledgements