{"id":190052,"date":"2024-05-25T07:24:31","date_gmt":"2024-05-25T12:24:31","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2024\/05\/these-two-scientists-have-mapped-out-the-insides-or-reachable-space-of-a-language-model-using-control-theory-what-they-discovered-was-extremely-surprising"},"modified":"2024-05-25T07:24:31","modified_gmt":"2024-05-25T12:24:31","slug":"these-two-scientists-have-mapped-out-the-insides-or-reachable-space-of-a-language-model-using-control-theory-what-they-discovered-was-extremely-surprising","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2024\/05\/these-two-scientists-have-mapped-out-the-insides-or-reachable-space-of-a-language-model-using-control-theory-what-they-discovered-was-extremely-surprising","title":{"rendered":"These two scientists have mapped out the insides or \u201creachable space\u201d of a language model using control theory, what they discovered was extremely surprising"},"content":{"rendered":"<p><\/p>\n<p><iframe style=\"display: block; margin: 0 auto; width: 100%; aspect-ratio: 4\/3; object-fit: contain;\" src=\"https:\/\/www.youtube.com\/embed\/Bpgloy1dDn0?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope;\n   picture-in-picture\" allowfullscreen><\/iframe><\/p>\n<p>Please support us on Patreon to get access to the private Discord server, bi-weekly calls, early access and ad-free listening.<br \/> \/ mlst.<\/p>\n<p>Aman Bhargava from Caltech and Cameron Witkowski from the University of Toronto to discuss their groundbreaking paper, \u201cWhat\u2019s the Magic Word? A Control Theory of LLM Prompting.\u201d They frame LLM systems as discrete stochastic dynamical systems. This means they look at LLMs in a structured way, similar to how we analyze control systems in engineering. They explore the \u201creachable set\u201d of outputs for an LLM. Essentially, this is the range of possible outputs the model can generate from a given starting point when influenced by different prompts. The research highlights that prompt engineering, or optimizing the input tokens, can significantly influence LLM outputs. They show that even short prompts can drastically alter the likelihood of specific outputs. Aman and Cameron\u2019s work might be a boon for understanding and improving LLMs. They suggest that a deeper exploration of control theory concepts could lead to more reliable and capable language models.<\/p>\n<p>What\u2019s the Magic Word? A Control Theory of LLM Prompting (Aman Bhargava, Cameron Witkowski, Manav Shah, Matt Thomson)<br \/>\n<a href=\"https:\/\/arxiv.org\/abs\/2310\">https:\/\/arxiv.org\/abs\/2310<\/a>.<\/p>\n<p>LLM Control Theory Seminar (April 2024)<br \/> \u2022 LLM Control Theory Seminar (April 2024) <\/p>\n<p>Society for the pursuit of AGI (Cameron founded it)<br \/>\n<a href=\"https:\/\/agisociety.mydurable.com\/\">https:\/\/agisociety.mydurable.com\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Please support us on Patreon to get access to the private Discord server, bi-weekly calls, early access and ad-free listening. \/ mlst. Aman Bhargava from Caltech and Cameron Witkowski from the University of Toronto to discuss their groundbreaking paper, \u201cWhat\u2019s the Magic Word? A Control Theory of LLM Prompting.\u201d They frame LLM systems as discrete [\u2026]<\/p>\n","protected":false},"author":661,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[13,6],"tags":[],"class_list":["post-190052","post","type-post","status-publish","format-standard","hentry","category-open-access","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/190052","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/661"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=190052"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/190052\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=190052"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=190052"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=190052"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}