spicysharkco/uncensored-leaderboard

Uncensored model leaderboard Public "uncensored" and abliterated models make the same claim: the refusals are gone and nothing else broke. This table measures both halves on one setup: how often each model still refuses harmful and harmless prompts, how far its outputs moved from the stock model, and what it scores on two task sets next to stock. One table per base model. Every row is pinned to a repo revision. Only scores are published here: no weights, model outputs or prompt… See the full description on the dataset page: https://huggingface.co/datasets/spicysharkco/uncensored-leaderboard.

<iframe src="https://silytics.si/embed/item/2fd8d19c-0d30-47eb-a0d6-771ff71a7e9c" 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 Information may be incomplete or out of date.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 20 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.