Remove intended-use and provenance sections from dataset card

README.md

@@ -98,22 +98,6 @@ with open("icbench.jsonl", encoding="utf-8") as f:
     records = [json.loads(line) for line in f if line.strip()]
 ```
 
-## Intended use and interpretation
-
-Use ICBench to evaluate whether a VLA policy responds to instruction semantics rather than relying only on visual task priors. Run each original and contradictory instruction in the corresponding, unchanged LIBERO environment. Successful completion of the original visually plausible task under a contradictory instruction represents instruction-inconsistent behavior; it should not be interpreted as correct instruction following. See the paper for the evaluation protocol and Linguistic Grounding Score.
-
-LIBERO must be installed separately to resolve the referenced BDDL paths and run simulation. The original instructions and paths are retained verbatim, including spelling and capitalization. These 30 simulated tasks and four perturbation categories provide a controlled diagnostic setting; they do not establish general real-world robot safety or comprehensive language understanding.
-
-## Provenance
-
-ICBench builds on [LIBERO](https://github.com/Lifelong-Robot-Learning/LIBERO). Please also acknowledge LIBERO when using its task definitions or simulator resources. This repository distributes the supplied `icbench.jsonl` instruction file; upstream software and assets retain their own licenses.
-
-Source-file SHA-256:
-
-```text
-8c253ead5175f8652ad931f56d370a1727bccc6f8b69162c80ef2319dad08630
-```
-
 ## License
 
 The ICBench instruction dataset is released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Please credit the authors and cite the paper when using this release. The full license text is in [`LICENSE`](LICENSE). LIBERO-derived material retains the upstream MIT notice in [`NOTICE`](NOTICE).