{"id":132631,"date":"2021-12-17T23:22:54","date_gmt":"2021-12-18T07:22:54","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2021\/12\/webgpt-improving-the-factual-accuracy-of-language-models-through-web-browsing"},"modified":"2021-12-17T23:22:54","modified_gmt":"2021-12-18T07:22:54","slug":"webgpt-improving-the-factual-accuracy-of-language-models-through-web-browsing","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2021\/12\/webgpt-improving-the-factual-accuracy-of-language-models-through-web-browsing","title":{"rendered":"WebGPT: Improving the factual accuracy of language models through web browsing"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/webgpt-improving-the-factual-accuracy-of-language-models-through-web-browsing3.jpg\"><\/a><\/p>\n<p>We\u2019ve fine-tuned GPT-3 to more accurately answer open-ended questions using a text-based web browser. Our prototype copies how humans research answers to questions online \u2013 it submits search queries, follows links, and scrolls up and down web pages. It is trained to cite its sources, which makes it easier to give feedback to improve factual accuracy. We\u2019re excited about developing more truthful AI, but challenges remain, such as coping with unfamiliar types of questions.<\/p>\n<p><a href=\"https:\/\/cdn.openai.com\/WebGPT.pdf\" class=\"\">Read paper<\/a><a href=\"https:\/\/openai.com\/blog\/improving-factual-accuracy\/#samples\" class=\"\">Browse samples<\/a><\/p>\n<p>Language models like GPT-3 are useful for many different tasks, but have a tendency to \u201challucinate\u201d information when performing tasks requiring obscure real-world knowledge. To address this, we taught GPT-3 to use a text-based web-browser. The model is provided with an open-ended question and a summary of the browser state, and must issue commands such as \u201cSearch \u2026\u201d, \u201cFind in page: \u2026\u201d or \u201cQuote: \u2026\u201d. In this way, the model collects passages from web pages, and then uses these to compose an answer.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>We\u2019ve fine-tuned GPT-3 to more accurately answer open-ended questions using a text-based web browser. Our prototype copies how humans research answers to questions online \u2013 it submits search queries, follows links, and scrolls up and down web pages. It is trained to cite its sources, which makes it easier to give feedback to improve factual [\u2026]<\/p>\n","protected":false},"author":359,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[],"class_list":["post-132631","post","type-post","status-publish","format-standard","hentry","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/132631","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\/359"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=132631"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/132631\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=132631"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=132631"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=132631"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}