{"id":3095623,"date":"2026-02-23T09:53:28","date_gmt":"2026-02-23T17:53:28","guid":{"rendered":"https:\/\/techcrunch.com\/?p=3095623"},"modified":"2026-02-23T13:27:24","modified_gmt":"2026-02-23T21:27:24","slug":"guide-labs-debuts-a-new-kind-of-interpretable-llm","status":"publish","type":"post","link":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","title":{"rendered":"Guide Labs debuts a new kind of interpretable LLM"},"content":{"rendered":"\n

The challenge of wrangling a deep learning model is often understanding why it does what it does: Whether it’s xAI’s repeated struggle sessions to fine-tune Grok’s odd politics, ChatGPT’s struggles with sycophancy, or run-of-the-mill hallucinations, plumbing through a neural network with billions of parameters isn’t easy.<\/p>\n\n

Guide Labs, a San Francisco startup founded by CEO Julius Adebayo and chief science officer Aya Abdelsalam Ismail, is offering an answer to that problem today. On Monday, the company open sourced an 8-billion-parameter LLM, Steerling-8B<\/a>, trained with a new architecture designed to make its actions easily interpretable: Every token produced by the model can be traced back to its origins in the LLM’s training data.<\/p>\n\n

That can be as simple as determining the reference materials for facts cited by the model, or as complex as understanding the model’s understanding of humor or gender.<\/p>\n\n

“If I have a trillion ways to encode gender, and I encode it in 1 billion of the 1 trillion things that I have, you have to make sure you find all those 1 billion things that I’ve encoded, and then you have to be able to reliably turn that on, turn them off,” Adebayo told TechCrunch. “You can do it with current models, but it’s very fragile\u00a0\u2026 It’s sort of one of the holy grail questions.”<\/p>\n\n

Adebayo began this work while earning his PhD at MIT, co-authoring a widely cited 2018 paper<\/a> that showed existing methods of understanding deep learning models were not reliable. That work ultimately led to the creation of a new way of building LLMs: Developers insert a concept layer in the model that buckets data into traceable categories. This requires more up-front data annotation, but by using other AI models to help, they were able to train this model as their largest proof of concept yet.<\/p>\n\n

“The kind of interpretability people do is\u00a0\u2026 neuroscience on a model, and we flip that,” Adebayo said. “What we do is actually engineer the model from the ground up so that you don’t need to do neuroscience.”<\/p>\n\n

\"\"
Image Credits:<\/strong>Guide Labs<\/span><\/figcaption><\/figure>\n\n

One concern with this approach is that it might eliminate some of the emergent behaviors that make LLMs so intriguing: Their ability to generalize in new ways about things they haven’t been trained on yet. Adebayo says that still happens in his company’s model: His team tracks what they call “discovered concepts” that the model discovered on its own, like quantum computing.<\/p>\n

\n\t
\n\t\t
\n\t\t\t<\/svg>\t\t<\/div>\n\t\t
Techcrunch event<\/div>\n\t\t
\n\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t

Save up to $300 or 30% to TechCrunch Founder Summit<\/h3>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t

1,000+ founders<\/strong> and investors come together at TechCrunch Founder Summit 2026<\/strong> for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately.

Offer ends March 13.<\/h4>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t

Save up to $300 or 30% to TechCrunch Founder Summit<\/h3>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t

1,000+ founders<\/strong> and investors come together at TechCrunch Founder Summit 2026<\/strong> for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately

Offer ends March 13.<\/h4>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t\t
\n\t\t\t\t\t\t\t\t\tBoston, MA<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t|<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\tJune 9, 2026<\/span>\n\t\t\t\t\t\t\t<\/div>\n\t\t\t
\n\t\t\t\t\n\n
\n
REGISTER NOW<\/a><\/div>\n<\/div>\n\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/div>\n<\/div>\n\n

Adebayo argues this interpretable architecture will be something everyone needs. For consumer-facing LLMs, these techniques should allow model builders to do things like block the use of copyrighted materials, or better control outputs around subjects like violence or drug abuse. Regulated industries will require more controllable LLMs \u2014 for example, in finance \u2014 where a model evaluating loan applicants needs to consider things like financial records but not race. There’s also a need for interpretability in scientific work, another area where Guide Labs has developed technology. Protein folding has been a big success for deep learning models, but scientists need more insight into why their software figured out promising combinations.<\/p>\n\n

“This model demonstrates that training interpretable models is no longer a sort of science; it’s now an engineering problem,” Adebayo said. “We figured out the science and we can scale them, and there is no reason why this kind of model wouldn’t match the performance of the frontier level models,” which have many more parameters.<\/p>\n\n

Guide Labs says that Steerling-8B can achieve 90% of the capability of existing models, but uses less training data, thanks to its novel architecture. The next step for the company, which emerged from Y Combinator and raised a $9 million seed round from Initialized Capital in November 2024, is to build a larger model and begin offering API and agentic access to users.<\/p>\n\n

“The way we’re currently training models is super primitive, and so democratizing inherent interpretability is actually going to be a long-term good thing for our role within the human race,” Adebayo told TechCrunch. “As we’re going after these models that are going to be super intelligent, you don’t want something to be making decisions on your behalf that’s sort of mysterious to you.”<\/p>\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.<\/p>\n","protected":false},"author":133574713,"featured_media":3095648,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"tc-featured-article":false,"tc-header-option":"","tc-breaking-news":false,"tc-article-brief":false,"carmot_uuid":"","apple_news_api_created_at":"2026-02-23T17:53:34Z","apple_news_api_id":"3c355c2a-6e49-4fc9-9504-dbace4cd4370","apple_news_api_modified_at":"2026-02-23T20:36:32Z","apple_news_api_revision":"AAAAAAAAAAAAAAAAAAAABQ==","apple_news_api_share_url":"https:\/\/apple.news\/APDVcKm5JT8mVBNus5M1DcA","apple_news_coverimage":0,"apple_news_coverimage_caption":"","apple_news_is_hidden":false,"apple_news_is_paid":false,"apple_news_is_preview":false,"apple_news_is_sponsored":false,"apple_news_maturity_rating":"","apple_news_metadata":"\"\"","apple_news_pullquote":"","apple_news_pullquote_position":"","apple_news_slug":"","apple_news_sections":[],"apple_news_suppress_video_url":false,"apple_news_use_image_component":false,"tc_subtitle":"","tc_featured_image_disabled":false,"tc_exclude_from_rss_feed":false,"tc_exclude_from_content_rivers":false,"footnotes":""},"categories":[577047203],"tags":[22376,577370565,173],"tc_region":[],"tc_event":[],"tc_storyline_tax":[],"coauthors":[577370422],"class_list":["post-3095623","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-artificial-intelligence","tag-exclusive","tag-guide-labs","tag-science"],"apple_news_notices":[],"yoast_head":"\nGuide Labs debuts a new kind of interpretable LLM | TechCrunch<\/title>\n<meta name=\"description\" content=\"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Guide Labs debuts a new kind of interpretable LLM | TechCrunch\" \/>\n<meta property=\"og:description\" content=\"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\" \/>\n<meta property=\"og:site_name\" content=\"TechCrunch\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/techcrunch\" \/>\n<meta property=\"article:published_time\" content=\"2026-02-23T17:53:28+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-02-23T21:27:24+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png?resize=1200,639\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"639\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Tim Fernholz\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@TechCrunch\" \/>\n<meta name=\"twitter:site\" content=\"@TechCrunch\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Tim Fernholz\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\"},\"author\":{\"name\":\"Tim Fernholz\",\"@id\":\"https:\/\/techcrunch.com\/#\/schema\/person\/5804b72644c60a6b5bf4ff8aefb59118\"},\"headline\":\"Guide Labs debuts a new kind of interpretable LLM\",\"datePublished\":\"2026-02-23T17:53:28+00:00\",\"dateModified\":\"2026-02-23T21:27:24+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\"},\"wordCount\":713,\"publisher\":{\"@id\":\"https:\/\/techcrunch.com\/#organization\"},\"image\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png\",\"keywords\":[\"Exclusive\",\"Guide Labs\",\"science\"],\"articleSection\":[\"AI\"],\"inLanguage\":\"en-US\",\"copyrightYear\":\"2026\",\"copyrightHolder\":{\"@id\":\"https:\/\/techcrunch.com\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\",\"url\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\",\"name\":\"Guide Labs debuts a new kind of interpretable LLM | TechCrunch\",\"isPartOf\":{\"@id\":\"https:\/\/techcrunch.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png\",\"datePublished\":\"2026-02-23T17:53:28+00:00\",\"dateModified\":\"2026-02-23T21:27:24+00:00\",\"description\":\"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.\",\"breadcrumb\":{\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage\",\"url\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png\",\"contentUrl\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png\",\"width\":1506,\"height\":802,\"caption\":\"Guide Labs' LLM traces its output back to training data.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/techcrunch.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Guide Labs debuts a new kind of interpretable LLM\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/techcrunch.com\/#website\",\"url\":\"https:\/\/techcrunch.com\/\",\"name\":\"TechCrunch\",\"description\":\"Startup and Technology News\",\"publisher\":{\"@id\":\"https:\/\/techcrunch.com\/#organization\"},\"alternateName\":\"TC\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/techcrunch.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/techcrunch.com\/#organization\",\"name\":\"TechCrunch\",\"alternateName\":\"TC\",\"url\":\"https:\/\/techcrunch.com\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/techcrunch.com\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2018\/04\/tc-logo-2018-square-reverse2x.png?resize=1200,1200\",\"contentUrl\":\"https:\/\/techcrunch.com\/wp-content\/uploads\/2018\/04\/tc-logo-2018-square-reverse2x.png?resize=1200,1200\",\"width\":1200,\"height\":1200,\"caption\":\"TechCrunch\"},\"image\":{\"@id\":\"https:\/\/techcrunch.com\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/techcrunch\",\"https:\/\/x.com\/TechCrunch\",\"https:\/\/mstdn.social\/@TechCrunch\",\"https:\/\/bsky.app\/profile\/techcrunch.com\",\"https:\/\/www.threads.net\/@techcrunch\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/techcrunch.com\/#\/schema\/person\/5804b72644c60a6b5bf4ff8aefb59118\",\"name\":\"Tim Fernholz\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/techcrunch.com\/#\/schema\/person\/image\/d6b7ee2d2792b9fe97616b183a691f6d\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/69fa76fc5ee34a5c21c897db2cdcba4089c942c99c34df4fd6eb586719f6298c?s=96&d=identicon&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/69fa76fc5ee34a5c21c897db2cdcba4089c942c99c34df4fd6eb586719f6298c?s=96&d=identicon&r=g\",\"caption\":\"Tim Fernholz\"},\"description\":\"Tim Fernholz is a journalist who writes about technology, finance and public policy. He has closely covered the rise of the private space industry and is the author of Rocket Billionaires: Elon Musk, Jeff Bezos and the New Space Race. Formerly, he was a senior reporter at Quartz, the global business news site, for more than a decade, and began his career as a political reporter in Washington, D.C. Now based in Oakland, California, he is cursed to live in interesting times. You can contact or verify outreach from Tim by emailing tim.fernholz@techcrunch.com or via an encrypted message to tim_fernholz.21 on Signal.\",\"sameAs\":[\"http:\/\/timfernholz.com\"],\"jobTitle\":\"Senior Reporter\",\"worksFor\":\"TechCrunch\",\"url\":\"https:\/\/techcrunch.com\/author\/tim-fernholz\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Guide Labs debuts a new kind of interpretable LLM | TechCrunch","description":"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.","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:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","og_locale":"en_US","og_type":"article","og_title":"Guide Labs debuts a new kind of interpretable LLM | TechCrunch","og_description":"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.","og_url":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","og_site_name":"TechCrunch","article_publisher":"https:\/\/www.facebook.com\/techcrunch","article_published_time":"2026-02-23T17:53:28+00:00","article_modified_time":"2026-02-23T21:27:24+00:00","og_image":[{"width":1200,"height":639,"url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png?resize=1200,639","type":"image\/png"}],"author":"Tim Fernholz","twitter_card":"summary_large_image","twitter_creator":"@TechCrunch","twitter_site":"@TechCrunch","twitter_misc":{"Written by":"Tim Fernholz","Est. reading time":"4 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#article","isPartOf":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/"},"author":{"name":"Tim Fernholz","@id":"https:\/\/techcrunch.com\/#\/schema\/person\/5804b72644c60a6b5bf4ff8aefb59118"},"headline":"Guide Labs debuts a new kind of interpretable LLM","datePublished":"2026-02-23T17:53:28+00:00","dateModified":"2026-02-23T21:27:24+00:00","mainEntityOfPage":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/"},"wordCount":713,"publisher":{"@id":"https:\/\/techcrunch.com\/#organization"},"image":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage"},"thumbnailUrl":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png","keywords":["Exclusive","Guide Labs","science"],"articleSection":["AI"],"inLanguage":"en-US","copyrightYear":"2026","copyrightHolder":{"@id":"https:\/\/techcrunch.com\/#organization"}},{"@type":"WebPage","@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","url":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","name":"Guide Labs debuts a new kind of interpretable LLM | TechCrunch","isPartOf":{"@id":"https:\/\/techcrunch.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage"},"image":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage"},"thumbnailUrl":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png","datePublished":"2026-02-23T17:53:28+00:00","dateModified":"2026-02-23T21:27:24+00:00","description":"The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.","breadcrumb":{"@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#primaryimage","url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png","contentUrl":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png","width":1506,"height":802,"caption":"Guide Labs' LLM traces its output back to training data."},{"@type":"BreadcrumbList","@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/techcrunch.com\/"},{"@type":"ListItem","position":2,"name":"Guide Labs debuts a new kind of interpretable LLM"}]},{"@type":"WebSite","@id":"https:\/\/techcrunch.com\/#website","url":"https:\/\/techcrunch.com\/","name":"TechCrunch","description":"Startup and Technology News","publisher":{"@id":"https:\/\/techcrunch.com\/#organization"},"alternateName":"TC","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/techcrunch.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/techcrunch.com\/#organization","name":"TechCrunch","alternateName":"TC","url":"https:\/\/techcrunch.com\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/techcrunch.com\/#\/schema\/logo\/image\/","url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2018\/04\/tc-logo-2018-square-reverse2x.png?resize=1200,1200","contentUrl":"https:\/\/techcrunch.com\/wp-content\/uploads\/2018\/04\/tc-logo-2018-square-reverse2x.png?resize=1200,1200","width":1200,"height":1200,"caption":"TechCrunch"},"image":{"@id":"https:\/\/techcrunch.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/techcrunch","https:\/\/x.com\/TechCrunch","https:\/\/mstdn.social\/@TechCrunch","https:\/\/bsky.app\/profile\/techcrunch.com","https:\/\/www.threads.net\/@techcrunch"]},{"@type":"Person","@id":"https:\/\/techcrunch.com\/#\/schema\/person\/5804b72644c60a6b5bf4ff8aefb59118","name":"Tim Fernholz","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/techcrunch.com\/#\/schema\/person\/image\/d6b7ee2d2792b9fe97616b183a691f6d","url":"https:\/\/secure.gravatar.com\/avatar\/69fa76fc5ee34a5c21c897db2cdcba4089c942c99c34df4fd6eb586719f6298c?s=96&d=identicon&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/69fa76fc5ee34a5c21c897db2cdcba4089c942c99c34df4fd6eb586719f6298c?s=96&d=identicon&r=g","caption":"Tim Fernholz"},"description":"Tim Fernholz is a journalist who writes about technology, finance and public policy. He has closely covered the rise of the private space industry and is the author of Rocket Billionaires: Elon Musk, Jeff Bezos and the New Space Race. Formerly, he was a senior reporter at Quartz, the global business news site, for more than a decade, and began his career as a political reporter in Washington, D.C. Now based in Oakland, California, he is cursed to live in interesting times. You can contact or verify outreach from Tim by emailing tim.fernholz@techcrunch.com or via an encrypted message to tim_fernholz.21 on Signal.","sameAs":["http:\/\/timfernholz.com"],"jobTitle":"Senior Reporter","worksFor":"TechCrunch","url":"https:\/\/techcrunch.com\/author\/tim-fernholz\/"}]}},"parsely":{"version":"1.1.0","canonical_url":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","smart_links":{"inbound":0,"outbound":0},"traffic_boost_suggestions_count":0,"meta":{"@context":"https:\/\/schema.org","@type":"NewsArticle","headline":"Guide Labs debuts a new kind of interpretable LLM","url":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/","mainEntityOfPage":{"@type":"WebPage","@id":"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/"},"thumbnailUrl":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png?w=150","image":{"@type":"ImageObject","url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png"},"articleSection":"AI","author":[{"@type":"Person","name":"Tim Fernholz"}],"creator":["Tim Fernholz"],"publisher":{"@type":"Organization","name":"TechCrunch","logo":"https:\/\/techcrunch.com\/wp-content\/uploads\/2015\/02\/cropped-cropped-favicon-gradient.png"},"keywords":["science","exclusive","guide labs"],"dateCreated":"2026-02-23T17:53:28Z","datePublished":"2026-02-23T17:53:28Z","dateModified":"2026-02-23T21:27:24Z"},"rendered":"<meta name=\"parsely-title\" content=\"Guide Labs debuts a new kind of interpretable LLM\" \/>\n<meta name=\"parsely-link\" content=\"https:\/\/techcrunch.com\/2026\/02\/23\/guide-labs-debuts-a-new-kind-of-interpretable-llm\/\" \/>\n<meta name=\"parsely-type\" content=\"post\" \/>\n<meta name=\"parsely-image-url\" content=\"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png?w=150\" \/>\n<meta name=\"parsely-pub-date\" content=\"2026-02-23T17:53:28Z\" \/>\n<meta name=\"parsely-section\" content=\"AI\" \/>\n<meta name=\"parsely-tags\" content=\"science,exclusive,guide labs\" \/>\n<meta name=\"parsely-author\" content=\"Tim Fernholz\" \/>","tracker_url":"https:\/\/cdn.parsely.com\/keys\/techcrunch.com\/p.js"},"jetpack_featured_media_url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2026\/02\/Screenshot-2026-02-23-at-9.31.05-AM.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/posts\/3095623","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/users\/133574713"}],"replies":[{"embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/comments?post=3095623"}],"version-history":[{"count":14,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/posts\/3095623\/revisions"}],"predecessor-version":[{"id":3095855,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/posts\/3095623\/revisions\/3095855"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/media\/3095648"}],"wp:attachment":[{"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/media?parent=3095623"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/categories?post=3095623"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/tags?post=3095623"},{"taxonomy":"tc_region","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/tc_region?post=3095623"},{"taxonomy":"tc_event","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/tc_event?post=3095623"},{"taxonomy":"tc_storyline_tax","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/tc_storyline_tax?post=3095623"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/techcrunch.com\/wp-json\/wp\/v2\/coauthors?post=3095623"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}