{"id":104482,"date":"2026-07-15T19:27:38","date_gmt":"2026-07-15T19:27:38","guid":{"rendered":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/"},"modified":"2026-07-15T19:27:38","modified_gmt":"2026-07-15T19:27:38","slug":"meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer","status":"publish","type":"post","link":"https:\/\/youzum.net\/ja\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/","title":{"rendered":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer"},"content":{"rendered":"<div data-chronoton-summary=\"&lt;ul&gt;&lt;br&gt;&lt;li&gt;&lt;strong&gt;An AI built to break AI:&lt;\/strong&gt; OpenAI created GPT-Red, an LLM trained specifically to hack other AI models. It automates &quot;red-teaming&quot;\u2014stress-testing systems for weaknesses\u2014a job that human testers alone can no longer keep pace with as AI grows more complex.&lt;\/li&gt;&lt;br&gt;&lt;li&gt;&lt;strong&gt;A sparring partner that gets smarter:&lt;\/strong&gt; GPT-Red was trained in a self-play loop, repeatedly attacking other models while they defended themselves. It outperformed human red-teamers at finding effective attacks and even discovered a previously unseen exploit called a &quot;fake chain of thought.&quot;&lt;\/li&gt;&lt;br&gt;&lt;li&gt;&lt;strong&gt;Real results, real limits:&lt;\/strong&gt; GPT-Red's training helped cut successful attacks on GPT-5.6. But it still struggles with conversational attacks and image-based exploits\u2014areas where human testers remain essential.&lt;\/li&gt;&lt;br&gt;&lt;\/ul&gt;\" data-chronoton-post-id=\"1140514\" data-chronoton-expand-collapse=\"1\" data-chronoton-analytics-enabled=\"1\"><\/div>\n<p>OpenAI has built an <a href=\"https:\/\/openai.com\/index\/unlocking-self-improvement-gpt-red\/\">LLM super-hacker called GPT-Red<\/a> that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its most robust release yet.<\/p>\n<p>GPT-Red automates a type of safety evaluation for software systems known as red-teaming, which is typically done by a team of human testers. The aim is to find as many different ways to break or hijack a system as possible. The weak spots can then be patched before the final version of the software is released.<\/p>\n<p>As LLMs become more complex and get used in a wider variety of tasks\u2014especially in the form of agents, which can interact with computer files, websites, and third-party code as well as other agents\u2014it\u2019s hard for teams of people by themselves to keep up with all the types of attacks that might take place. \u201cThe risk surface grows and the blast radius also grows,\u201d says Nikhil Kandpal, a research scientist at OpenAI who co-created GPT-Red.<\/p>\n<p>OpenAI built GPT-Red to future-proof its safety testing process. \u201cAs more capable models become available, we will have already designed the system that can discover new modes of attack,\u201d says Dylan Hunn, a research scientist at the company and fellow co-creator of GPT-Red. The researchers say it has already come up with new types of attack that had not been seen before.<\/p>\n<p>OpenAI focused most of its efforts on a type of attack known as a prompt injection, where a hacker slips an LLM instructions to make it do things its developers or users do not want it to, such as copy confidential information, sabotage a company\u2019s code base, or generate embarrassing or harmful output. In theory, such instructions can be hidden in any text that the LLM might encounter\u2014in code or on a website, for example.\u00a0\u00a0\u00a0\u00a0<\/p>\n<h3 class=\"wp-block-heading\"><strong>Training dojo<\/strong><\/h3>\n<p>To build GPT-Red, OpenAI\u2019s researchers took an LLM that had not been trained as a hacker and set it up in what\u2019s known as a self-play loop with several other models. Its goal was to try to attack the other models; their goal was to try to defend themselves. Over many rounds of play, GPT-Red became better and better at attacking other LLMs, and those LLMs became better and better at fending off the attacks.<\/p>\n<p>The training took place in a kind of dojo that OpenAI had designed to mimic a range of scenarios in which LLMs might be deployed in the real world, including browsing the web, reading emails or calendar apps, and editing code.\u00a0\u00a0<\/p>\n<p>When GPT-Red found a new kind of attack, it would explore multiple different versions of it to find the most efficient one for specific scenarios. \u201cCompared to a human red-teamer, the model is very, very good at finding exactly what will work, exactly what\u2019s most effective,\u201d says Hunn. \u201cIt\u2019s extremely persistent about drilling down into an attack that it has discovered.\u201d\u00a0\u00a0<\/p>\n<p>In particular, OpenAI claims that GPT-Red found a type of prompt injection attack that the researchers had not seen before, which they call a fake chain of thought. A chain of thought is a kind of diary in which an LLM makes notes to itself and keeps track of partial results as it works through problems. GPT-Red found a way to insert a fake entry into another model\u2019s chain of thought that would trick that model into acting on spoofed information.<\/p>\n<p>\u201cIt\u2019s like if I told you that 1+1=3 and that you have verified this already,\u201d says Chris Choquette-Choo, another research scientist on the team. \u201cThe model\u2019s like, \u2018Oh, okay, of course,\u2019 and it just spits out 3.\u201d<\/p>\n<p>Jessica Ji, a senior research analyst who works on AI security at Georgetown University\u2019s Center for Security and Emerging Technology (CSET), thinks the self-play loop that OpenAI used is a good approach. \u201cThe results look very promising,\u201d she says.<\/p>\n<p>OpenAI tested how good an attacker GPT-Red was by rerunning an experiment from 2025 in which human red-teamers tried to find weaknesses in an earlier version of GPT-5. When GPT-Red was set the same task, it was more successful at finding effective attacks than the humans had been.<\/p>\n<p>OpenAI also tested GPT-Red against Vendy, a vending machine agent developed by Andon Labs, a company that assesses how well agents perform real-world tasks. GPT-Red was able to hack Vendy to make it change the prices of items on sale and cancel a customer\u2019s order.<\/p>\n<h3 class=\"wp-block-heading\"><strong>Defensive behavior<\/strong><\/h3>\n<p>OpenAI says that when it tried out some of the strongest attacks that GPT-Red had come up with on its models, more than 90% of them worked against GPT-5 (released in August last year), and fewer than 23% worked against the new GPT-5.6.<\/p>\n<p>GPT-Red isn\u2019t perfect. It is not great at figuring out attacks that involve a back-and-forth conversation between hacker and target, something that human attackers would have few problems with. It is also not yet that great at using images, which can be used to pass text to models in prompt injection attacks.\u00a0\u00a0\u00a0\u00a0<\/p>\n<p>The company says that GPT-Red supplements the work of its human red-teamers; people can find attacks it misses, and vice versa. One approach OpenAI is taking is to give GPT-Red an attack that humans came up with and ask it to find all the variations.<\/p>\n<p>\u201cI think human expertise will still be very important,\u201d says CSET\u2019s Ji. \u201cIt would be really useful to be able to distinguish where human testing is most needed.\u201d<\/p>\n<p>Unsurprisingly, OpenAI will not be releasing GPT-Red. The company is also confident that the super-hacker is stronger than any copycat model someone might try to create. The researchers say they have been working on the model for more than a year, backed by the compute resources of one of the richest companies in the world.<\/p>\n<p>\u201cIt\u2019s not a trivial thing that someone else could easily do\u2014you know, just go and train a super-attacker using this idea,\u201d says Choquette-Choo.<\/p>","protected":false},"excerpt":{"rendered":"<p>OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its most robust release yet. GPT-Red automates a type of safety evaluation for software systems known as red-teaming, which is typically done by a team of human testers. The aim is to find as many different ways to break or hijack a system as possible. The weak spots can then be patched before the final version of the software is released. As LLMs become more complex and get used in a wider variety of tasks\u2014especially in the form of agents, which can interact with computer files, websites, and third-party code as well as other agents\u2014it\u2019s hard for teams of people by themselves to keep up with all the types of attacks that might take place. \u201cThe risk surface grows and the blast radius also grows,\u201d says Nikhil Kandpal, a research scientist at OpenAI who co-created GPT-Red. OpenAI built GPT-Red to future-proof its safety testing process. \u201cAs more capable models become available, we will have already designed the system that can discover new modes of attack,\u201d says Dylan Hunn, a research scientist at the company and fellow co-creator of GPT-Red. The researchers say it has already come up with new types of attack that had not been seen before. OpenAI focused most of its efforts on a type of attack known as a prompt injection, where a hacker slips an LLM instructions to make it do things its developers or users do not want it to, such as copy confidential information, sabotage a company\u2019s code base, or generate embarrassing or harmful output. In theory, such instructions can be hidden in any text that the LLM might encounter\u2014in code or on a website, for example.\u00a0\u00a0\u00a0\u00a0 Training dojo To build GPT-Red, OpenAI\u2019s researchers took an LLM that had not been trained as a hacker and set it up in what\u2019s known as a self-play loop with several other models. Its goal was to try to attack the other models; their goal was to try to defend themselves. Over many rounds of play, GPT-Red became better and better at attacking other LLMs, and those LLMs became better and better at fending off the attacks. The training took place in a kind of dojo that OpenAI had designed to mimic a range of scenarios in which LLMs might be deployed in the real world, including browsing the web, reading emails or calendar apps, and editing code.\u00a0\u00a0 When GPT-Red found a new kind of attack, it would explore multiple different versions of it to find the most efficient one for specific scenarios. \u201cCompared to a human red-teamer, the model is very, very good at finding exactly what will work, exactly what\u2019s most effective,\u201d says Hunn. \u201cIt\u2019s extremely persistent about drilling down into an attack that it has discovered.\u201d\u00a0\u00a0 In particular, OpenAI claims that GPT-Red found a type of prompt injection attack that the researchers had not seen before, which they call a fake chain of thought. A chain of thought is a kind of diary in which an LLM makes notes to itself and keeps track of partial results as it works through problems. GPT-Red found a way to insert a fake entry into another model\u2019s chain of thought that would trick that model into acting on spoofed information. \u201cIt\u2019s like if I told you that 1+1=3 and that you have verified this already,\u201d says Chris Choquette-Choo, another research scientist on the team. \u201cThe model\u2019s like, \u2018Oh, okay, of course,\u2019 and it just spits out 3.\u201d Jessica Ji, a senior research analyst who works on AI security at Georgetown University\u2019s Center for Security and Emerging Technology (CSET), thinks the self-play loop that OpenAI used is a good approach. \u201cThe results look very promising,\u201d she says. OpenAI tested how good an attacker GPT-Red was by rerunning an experiment from 2025 in which human red-teamers tried to find weaknesses in an earlier version of GPT-5. When GPT-Red was set the same task, it was more successful at finding effective attacks than the humans had been. OpenAI also tested GPT-Red against Vendy, a vending machine agent developed by Andon Labs, a company that assesses how well agents perform real-world tasks. GPT-Red was able to hack Vendy to make it change the prices of items on sale and cancel a customer\u2019s order. Defensive behavior OpenAI says that when it tried out some of the strongest attacks that GPT-Red had come up with on its models, more than 90% of them worked against GPT-5 (released in August last year), and fewer than 23% worked against the new GPT-5.6. GPT-Red isn\u2019t perfect. It is not great at figuring out attacks that involve a back-and-forth conversation between hacker and target, something that human attackers would have few problems with. It is also not yet that great at using images, which can be used to pass text to models in prompt injection attacks.\u00a0\u00a0\u00a0\u00a0 The company says that GPT-Red supplements the work of its human red-teamers; people can find attacks it misses, and vice versa. One approach OpenAI is taking is to give GPT-Red an attack that humans came up with and ask it to find all the variations. \u201cI think human expertise will still be very important,\u201d says CSET\u2019s Ji. \u201cIt would be really useful to be able to distinguish where human testing is most needed.\u201d Unsurprisingly, OpenAI will not be releasing GPT-Red. The company is also confident that the super-hacker is stronger than any copycat model someone might try to create. The researchers say they have been working on the model for more than a year, backed by the compute resources of one of the richest companies in the world. \u201cIt\u2019s not a trivial thing that someone else could easily do\u2014you know, just go and train a super-attacker using this idea,\u201d says Choquette-Choo.<\/p>","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"pmpro_default_level":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_pvb_checkbox_block_on_post":false,"footnotes":""},"categories":[52,5,7,1],"tags":[],"class_list":["post-104482","post","type-post","status-publish","format-standard","hentry","category-ai-club","category-committee","category-news","category-uncategorized","pmpro-has-access"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v25.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum<\/title>\n<meta name=\"description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/youzum.net\/ja\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\" \/>\n<meta property=\"og:locale\" content=\"ja_JP\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum\" \/>\n<meta property=\"og:description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta property=\"og:url\" content=\"https:\/\/youzum.net\/ja\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\" \/>\n<meta property=\"og:site_name\" content=\"YouZum\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/DroneAssociationTH\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-07-15T19:27:38+00:00\" \/>\n<meta name=\"author\" content=\"admin NU\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u57f7\u7b46\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin NU\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u63a8\u5b9a\u8aad\u307f\u53d6\u308a\u6642\u9593\" \/>\n\t<meta name=\"twitter:data2\" content=\"5\u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\"},\"author\":{\"name\":\"admin NU\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\"},\"headline\":\"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer\",\"datePublished\":\"2026-07-15T19:27:38+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\"},\"wordCount\":1023,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"articleSection\":[\"AI\",\"Committee\",\"News\",\"Uncategorized\"],\"inLanguage\":\"ja\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\",\"url\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\",\"name\":\"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum\",\"isPartOf\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#website\"},\"datePublished\":\"2026-07-15T19:27:38+00:00\",\"description\":\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\",\"breadcrumb\":{\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#breadcrumb\"},\"inLanguage\":\"ja\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/youzum.net\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/yousum.gpucore.co\/#website\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"name\":\"YouSum\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/yousum.gpucore.co\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"ja\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\",\"name\":\"Drone Association Thailand\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"ja\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"width\":300,\"height\":300,\"caption\":\"Drone Association Thailand\"},\"image\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/DroneAssociationTH\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\",\"name\":\"admin NU\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"ja\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"caption\":\"admin NU\"},\"url\":\"https:\/\/youzum.net\/ja\/members\/adminnu\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/youzum.net\/ja\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/","og_locale":"ja_JP","og_type":"article","og_title":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum","og_description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","og_url":"https:\/\/youzum.net\/ja\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/","og_site_name":"YouZum","article_publisher":"https:\/\/www.facebook.com\/DroneAssociationTH\/","article_published_time":"2026-07-15T19:27:38+00:00","author":"admin NU","twitter_card":"summary_large_image","twitter_misc":{"\u57f7\u7b46\u8005":"admin NU","\u63a8\u5b9a\u8aad\u307f\u53d6\u308a\u6642\u9593":"5\u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#article","isPartOf":{"@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/"},"author":{"name":"admin NU","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c"},"headline":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer","datePublished":"2026-07-15T19:27:38+00:00","mainEntityOfPage":{"@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/"},"wordCount":1023,"commentCount":0,"publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"articleSection":["AI","Committee","News","Uncategorized"],"inLanguage":"ja","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/","url":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/","name":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer - YouZum","isPartOf":{"@id":"https:\/\/yousum.gpucore.co\/#website"},"datePublished":"2026-07-15T19:27:38+00:00","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","breadcrumb":{"@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#breadcrumb"},"inLanguage":"ja","potentialAction":[{"@type":"ReadAction","target":["https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/youzum.net\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/youzum.net\/"},{"@type":"ListItem","position":2,"name":"Meet GPT-Red: an LLM super-hacker OpenAI built to make its models safer"}]},{"@type":"WebSite","@id":"https:\/\/yousum.gpucore.co\/#website","url":"https:\/\/yousum.gpucore.co\/","name":"YouSum","description":"","publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/yousum.gpucore.co\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"ja"},{"@type":"Organization","@id":"https:\/\/yousum.gpucore.co\/#organization","name":"Drone Association Thailand","url":"https:\/\/yousum.gpucore.co\/","logo":{"@type":"ImageObject","inLanguage":"ja","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","width":300,"height":300,"caption":"Drone Association Thailand"},"image":{"@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/DroneAssociationTH\/"]},{"@type":"Person","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c","name":"admin NU","image":{"@type":"ImageObject","inLanguage":"ja","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","caption":"admin NU"},"url":"https:\/\/youzum.net\/ja\/members\/adminnu\/"}]}},"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"admin NU","author_link":"https:\/\/youzum.net\/ja\/members\/adminnu\/"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/youzum.net\/ja\/category\/ai-club\/\" rel=\"category tag\">AI<\/a> <a href=\"https:\/\/youzum.net\/ja\/category\/committee\/\" rel=\"category tag\">Committee<\/a> <a href=\"https:\/\/youzum.net\/ja\/category\/news\/\" rel=\"category tag\">News<\/a> <a href=\"https:\/\/youzum.net\/ja\/category\/uncategorized\/\" rel=\"category tag\">Uncategorized<\/a>","rttpg_excerpt":"OpenAI has built an LLM super-hacker called GPT-Red that it uses as a sparring partner to help its other models boost their defenses against cyberattacks. Last week the company released the latest version of its flagship LLM, GPT-5.6. OpenAI says that training it against GPT-Red made the model its most robust release yet. GPT-Red automates&hellip;","_links":{"self":[{"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/posts\/104482","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/comments?post=104482"}],"version-history":[{"count":0,"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/posts\/104482\/revisions"}],"wp:attachment":[{"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/media?parent=104482"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/categories?post=104482"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/youzum.net\/ja\/wp-json\/wp\/v2\/tags?post=104482"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}