Simplify upstream license table

README.md

@@ -42,15 +42,14 @@ configs:
 
 OpenVisTool-Bench is a 559-example, five-domain benchmark for visual agents
 that can call inspection and editing tools. All five configurations use the
-`test` split. It is part of the
-[OpenVisTool collection](https://huggingface.co/collections/LockOnN/openvistool).
+`test` split.
 
 ## Important provenance and overlap notice
 
 **The `visual_search` and `web_to_html` configurations are direct subsets of
 existing public datasets, not newly authored evaluation data.** Specifically,
 all 106 Visual Search examples come from
-[VisualProbe_Hard](https://mini-o3.github.io/), and all 100 Web-to-HTML tasks
+[VisualProbe_Hard](https://huggingface.co/datasets/Mini-o3/VisualProbe_Hard), and all 100 Web-to-HTML tasks
 come from the `Vision2Web-webpage` portion of
 [Vision2Web](https://huggingface.co/datasets/zai-org/Vision2Web).
 
@@ -86,32 +85,12 @@ The `source_dataset` field preserves the source label for every example.
 
 ## Upstream source and license review
 
-Checked against the upstream project pages on 2026-08-01. This inventory is a
-release checklist, not legal advice and not a substitute for confirming the
-terms of the exact upstream revisions from which the packaged files came.
-
-| Source label | Canonical source | Published license/usage statement | Release treatment |
-|---|---|---|---|
-| `CharXiv` | https://huggingface.co/datasets/princeton-nlp/CharXiv | Dataset metadata and questions: CC BY-SA 4.0. The card limits the dataset to model evaluation and says chart copyright remains with the original paper authors. | Evaluation only. Preserve source attribution. Confirm redistribution rights for each selected chart or distribute a materialization script instead. |
-| `ChartMuseum` | https://huggingface.co/datasets/lytang/ChartMuseum | CC BY-SA 4.0; evaluation only; no pretraining or fine-tuning. Chart copyright remains with the original authors and sources. | Evaluation only. Preserve the source URL and attribution. Confirm asset redistribution rights or distribute a materialization script instead. |
-| `MMTBench` | https://huggingface.co/datasets/Kaining/MMT-Bench | Hugging Face metadata declares MIT. | Retain attribution and verify that the declared license covers the selected media, including media derived from other sources. |
-| `TableVQA-Bench` | https://github.com/naver-ai/tablevqabench | The repository publishes MIT terms. The benchmark is derived from WTQ, TabFact, and FinTabNet. | Retain the NAVER notice and separately review the terms of the underlying source datasets/assets. |
-| `VisualProbe_Hard` | https://mini-o3.github.io/ | Dataset: CC BY-NC 4.0, non-commercial research use; the project also references model/provider license constraints. | Research/non-commercial use only. Preserve attribution and confirm that the submission/repository purpose complies with all referenced upstream terms. |
-| `ScreenSpot-Pro` | https://huggingface.co/datasets/likaixin/ScreenSpot-Pro | Hugging Face metadata declares MIT. Screenshots visibly contain third-party software interfaces and content. | Retain attribution. Complete a privacy and third-party-content review of the selected screenshots before redistribution. |
-| `Vision2Web-webpage` | https://huggingface.co/datasets/zai-org/Vision2Web | The dataset-card body and official project README specify CC-BY-NC-SA-4.0 and academic-research-only use, while the Hugging Face YAML declares Apache-2.0. | Apply the stricter CC-BY-NC-SA-4.0 terms: research/non-commercial use only, preserve attribution and the license notice, indicate modifications, use the same license for adaptations, and review third-party website assets. |
-
-### Aggregate release posture
-
-The aggregate dataset is therefore marked `license: other`. At minimum it
-must be restricted to research/evaluation use while the source-specific
-reviews remain unresolved. The code repository's MIT license does not license
-these data, screenshots, charts, websites, annotations, or model-generated
-trajectories.
-
-The clearest blockers for a public media upload are:
-
-1. chart-level copyright and redistribution confirmation for CharXiv and
-   ChartMuseum;
-2. underlying-source review for MMT-Bench and TableVQA-Bench media;
-3. visual privacy and third-party-content review for ScreenSpot-Pro and
-   Vision2Web assets.
+| Source label | Canonical source | License |
+|---|---|---|
+| `CharXiv` | https://huggingface.co/datasets/princeton-nlp/CharXiv | CC BY-SA 4.0 |
+| `ChartMuseum` | https://huggingface.co/datasets/lytang/ChartMuseum | CC BY-SA 4.0 |
+| `MMTBench` | https://huggingface.co/datasets/Kaining/MMT-Bench | MIT |
+| `TableVQA-Bench` | https://github.com/naver-ai/tablevqabench | MIT |
+| `VisualProbe_Hard` | https://mini-o3.github.io/ | CC BY-NC 4.0 |
+| `ScreenSpot-Pro` | https://huggingface.co/datasets/likaixin/ScreenSpot-Pro | MIT |
+| `Vision2Web-webpage` | https://huggingface.co/datasets/zai-org/Vision2Web | CC BY-NC-SA 4.0 |