{"id":216172,"date":"2025-06-18T09:12:16","date_gmt":"2025-06-18T14:12:16","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2025\/06\/lost-in-the-middle-how-llm-architecture-and-training-data-shape-ais-position-bias"},"modified":"2025-06-18T09:12:16","modified_gmt":"2025-06-18T14:12:16","slug":"lost-in-the-middle-how-llm-architecture-and-training-data-shape-ais-position-bias","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2025\/06\/lost-in-the-middle-how-llm-architecture-and-training-data-shape-ais-position-bias","title":{"rendered":"Lost in the middle: How LLM architecture and training data shape AI\u2019s position bias"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/lost-in-the-middle-how-llm-architecture-and-training-data-shape-ais-position-bias2.jpg\"><\/a><\/p>\n<p>Research has shown that large language models (LLMs) tend to overemphasize information at the beginning and end of a document or conversation, while neglecting the middle.<\/p>\n<p>This \u201cposition bias\u201d means that if a lawyer is using an LLM-powered virtual assistant to retrieve a certain phrase in a 30-page affidavit, the LLM is more likely to find the right text if it is on the initial or final pages.<\/p>\n<p>MIT researchers have discovered the mechanism behind this phenomenon.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Research has shown that large language models (LLMs) tend to overemphasize information at the beginning and end of a document or conversation, while neglecting the middle. This \u201cposition bias\u201d means that if a lawyer is using an LLM-powered virtual assistant to retrieve a certain phrase in a 30-page affidavit, the LLM is more likely to [\u2026]<\/p>\n","protected":false},"author":427,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[],"class_list":["post-216172","post","type-post","status-publish","format-standard","hentry","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/216172","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\/427"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=216172"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/216172\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=216172"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=216172"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=216172"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}