{"id":5189,"date":"2024-03-10T09:19:29","date_gmt":"2024-03-10T01:19:29","guid":{"rendered":"https:\/\/www.1ai.net\/?p=5189"},"modified":"2024-03-10T09:19:29","modified_gmt":"2024-03-10T01:19:29","slug":"%e5%8f%af%e6%a3%80%e6%b5%8b-ai%e6%a8%a1%e5%9e%8b%e4%b8%ad%e7%89%88%e6%9d%83%e5%86%85%e5%ae%b9%ef%bc%8cpatronus-%e6%8e%a8%e5%87%ba-copyrightcatcher-api","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/5189.html","title":{"rendered":"Patronus launches CopyrightCatcher API to detect copyrighted content in AI models"},"content":{"rendered":"<p data-vmark=\"5b67\">Specially developed<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%a4%a7%e8%af%ad%e8%a8%80%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [large language model]]\" target=\"_blank\" >Large Language Model<\/a>(LLM) Assessment Tool\u00a0<a href=\"https:\/\/www.1ai.net\/en\/tag\/patronus-ai\" title=\"_Other Organiser\" target=\"_blank\" >Patronus AI<\/a> Recently, an API called &quot;CopyrightCatcher&quot; was released, which can be used to detect whether the output results of large language models contain copyrights.<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e4%be%b5%e6%9d%83%e5%86%85%e5%ae%b9\" title=\"[Sees articles with [tort content] labels]\" target=\"_blank\" >Infringing Content<\/a>, the relevant tool DEMO has been released, interested friends can<a href=\"https:\/\/copyrightcatcher.patronus.ai\/\" target=\"_blank\" rel=\"noopener\">Click here to visit<\/a>download.<\/p>\n<p data-vmark=\"15fd\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5190\" title=\"f8a7aec7-0225-4153-9c88-1af0fcced3a4\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/03\/f8a7aec7-0225-4153-9c88-1af0fcced3a4.png\" alt=\"f8a7aec7-0225-4153-9c88-1af0fcced3a4\" width=\"1440\" height=\"890\" \/><\/p>\n<p>\u25b2 Image source: Patronus AI official press release<\/p>\n<p data-vmark=\"9fee\">Patronus AI said that the training data of the common large language models on the market often contains copyrighted content, so these models can easily output corresponding copyrighted content, which brings significant legal risks to companies that deploy related models. Therefore, they launched the CopyrightCatcher API to solve related infringement issues.<\/p>\n<p data-vmark=\"9a20\">According to reports, in order to check whether the output data of the large language model contains infringing content, Patronus AI researchers extracted a batch of copyrighted text samples from the Goodreads book platform to conduct adversarial training on the model.<span class=\"accentTextColor\">Based on these books, 100 suggestive passages were created.<\/span>.<\/p>\n<p data-vmark=\"fbdf\">According to the report, 50 of the relevant passages require the model to &quot;generate the first paragraph of the book&quot;, and the other 50 require the model to generate text fragments in the book. The researchers compiled and compiled the above passages into the CopyrightCatcher API.<span class=\"accentTextColor\">It claims to be able to detect how large language models &quot;precisely copy content from the original training data&quot; and also assess the probability of the model outputting infringing content.<\/span>.<\/p>\n<p data-vmark=\"93c8\">The researchers used OpenAI&#039;s GPT-4, Mistral&#039;s Mixtral-8x7B-Instruct-v0.1, Anthropic&#039;s Claude-2.1, and Meta&#039;s Llama-2-70b-chat for testing.<span class=\"accentTextColor\">It was finally found that GPT-4 was most likely to generate infringing content, while Claude-2.1 was the least likely to generate infringing content.<\/span>:<\/p>\n<ul class=\"list-paddingleft-2\">\n<li>\n<p data-vmark=\"dad2\">GPT-4: 44%<\/p>\n<\/li>\n<li>\n<p data-vmark=\"e3f5\">Mixtral-8x7B-Instruct-v0.1:22%<\/p>\n<\/li>\n<li>\n<p data-vmark=\"fd75\">Llama-2-70b-chat:10%<\/p>\n<\/li>\n<li>\n<p data-vmark=\"d636\">Claude-2.1:8%<\/p>\n<\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>Patronus AI, which specializes in the development of large language model (LLM) evaluation tools, has recently released an API called \"CopyrightCatcher\", which can be used to detect whether the output of LLM contains infringing content, and the DEMO of the relevant tool has been released. The DEMO of the tool has been released, so if you are interested, you can click here to download it. \u25b2Patronus AI official press release Patronus AI said that the training data of common big language models in the market often contains copyrighted content, so it is easy for these models to output copyrighted content, which poses a significant legal risk to organizations deploying these models, and therefore they have launched the CopyrightCatcher API.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[1588,1589,706],"collection":[],"class_list":["post-5189","post","type-post","status-publish","format-standard","hentry","category-news","tag-patronus-ai","tag-1589","tag-706"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5189","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=5189"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5189\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=5189"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=5189"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=5189"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=5189"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}