List canonical LexBench datasets and original Headless Browser inputs
Browse files
README.md
CHANGED
|
@@ -21,13 +21,14 @@ Our versioned benchmark datasets are available here:
|
|
| 21 |
|
| 22 |
| Dataset | What it evaluates |
|
| 23 |
| --- | --- |
|
| 24 |
-
| [LexBench
|
| 25 |
-
| [LexBench
|
| 26 |
-
| [LexBench
|
| 27 |
-
| [LexBench
|
| 28 |
-
| [LexBench
|
| 29 |
-
| [LexBench
|
| 30 |
-
| [LexBench
|
|
|
|
| 31 |
|
| 32 |
Each dataset card describes its task scope, use, sources and evaluation limits. Derived datasets retain upstream attribution and document their changes.
|
| 33 |
|
|
|
|
| 21 |
|
| 22 |
| Dataset | What it evaluates |
|
| 23 |
| --- | --- |
|
| 24 |
+
| [LexBench-Web-Fetch](https://huggingface.co/datasets/Lexmount/LexBench-Web-Fetch) | Page content, documents, structured records, resource discovery and file delivery |
|
| 25 |
+
| [LexBench-Web-Search](https://huggingface.co/datasets/Lexmount/LexBench-Web-Search) | Search relevance, ranking and explicit source requirements |
|
| 26 |
+
| [LexBench-WebMainBench](https://huggingface.co/datasets/Lexmount/LexBench-WebMainBench) | Main-content extraction from frozen HTML, using a reference-corrected upstream edition |
|
| 27 |
+
| [LexBench-SimpleQA-Verified](https://huggingface.co/datasets/Lexmount/LexBench-SimpleQA-Verified) | Reference-page retrieval and answer-bearing results for unchanged publisher questions selected by length |
|
| 28 |
+
| [LexBench-WCXB](https://huggingface.co/datasets/Lexmount/LexBench-WCXB) | Source-qualified frozen web pages and extraction references |
|
| 29 |
+
| [LexBench-Headless-Browser](https://huggingface.co/datasets/Lexmount/LexBench-Headless-Browser) | Raw CDP, browser-driver compatibility and web-platform behavior on deterministic fixtures |
|
| 30 |
+
| [LexBench-Trajectory-Replay](https://huggingface.co/datasets/Lexmount/LexBench-Trajectory-Replay) | Curated fixed-action browser tasks and replay trajectories |
|
| 31 |
+
| [LexBench-WebArena-Replay](https://huggingface.co/datasets/Lexmount/LexBench-WebArena-Replay) | Separate human-recording, corrected-task and full official-task configurations |
|
| 32 |
|
| 33 |
Each dataset card describes its task scope, use, sources and evaluation limits. Derived datasets retain upstream attribution and document their changes.
|
| 34 |
|