keenanpepper/arc-whestbench-higher-moments-2026

WhestBench Higher Cross-Moments — pre & post activation ground truth Monte-Carlo ground-truth higher moments of the activations of the 1000 random ReLU MLPs in the ARC WhiteBox Estimation Challenge 2026 full split. The official challenge dataset (aicrowd/arc-whestbench-public-2026) ships only the per-neuron mean post-activation. This dataset adds the full set of O(W²) moment-tensor components up to 4th order, for both the pre-activation and the post-activation at every layer —… See the full description on the dataset page: https://huggingface.co/datasets/keenanpepper/arc-whestbench-higher-mome

<iframe src="https://silytics.si/embed/item/6ea6d274-c650-4798-a983-bd079ce603d4" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.