{"id":41918,"date":"2026-07-30T01:27:12","date_gmt":"2026-07-29T23:27:12","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/what-are-the-best-publicly-available-uncensored-datasets\/"},"modified":"2026-07-30T01:27:12","modified_gmt":"2026-07-29T23:27:12","slug":"what-are-the-best-publicly-available-uncensored-datasets","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/what-are-the-best-publicly-available-uncensored-datasets\/","title":{"rendered":"What Are The Best Publicly Available &#8220;uncensored&#8221; Datasets?"},"content":{"rendered":"<p><!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>I use &#8220;Heretic&#8221; library on models to liberate them from their safeguards, but while checking their &#8220;uncensoredness&#8221;, I found they can hallucinate a lot. You know, it&#8217;s basically like a child who&#8217;s now allowed to use the F word once and he says &#8220;Fred&#8221; instead of the actual thing. <\/p>\n<p>So I think if the models train on valid uncensored data (specially if they start Grokking) the results can improve. So I am using for these types of datasets to test my theory. <\/p>\n<\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/Haghiri75\"> \/u\/Haghiri75 <\/a> <br \/> <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1vabhfw\/what_are_the_best_publicly_available_uncensored\/\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1vabhfw\/what_are_the_best_publicly_available_uncensored\/\">[comments]<\/a><\/span><\/p><div class='watch-action'><div class='watch-position align-right'><div class='action-like'><a class='lbg-style1 like-41918 jlk' href='javascript:void(0)' data-task='like' data-post_id='41918' data-nonce='3bde7dc5cf' rel='nofollow'><img class='wti-pixel' src='https:\/\/www.graviton.at\/letterswaplibrary\/wp-content\/plugins\/wti-like-post\/images\/pixel.gif' title='Like' \/><span class='lc-41918 lc'>0<\/span><\/a><\/div><\/div> <div class='status-41918 status align-right'><\/div><\/div><div class='wti-clear'><\/div>","protected":false},"excerpt":{"rendered":"<p>I use &#8220;Heretic&#8221; library on models to liberate them from their safeguards, but while checking their &#8220;uncensoredness&#8221;,&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[85],"tags":[],"class_list":["post-41918","post","type-post","status-publish","format-standard","hentry","category-datatards","wpcat-85-id"],"_links":{"self":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41918","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/comments?post=41918"}],"version-history":[{"count":0,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41918\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/media?parent=41918"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/categories?post=41918"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/tags?post=41918"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}