{"id":2815,"date":"2026-06-02T18:02:33","date_gmt":"2026-06-02T16:02:33","guid":{"rendered":"https:\/\/www.visual61.com\/site\/?p=2815"},"modified":"2026-07-17T21:55:29","modified_gmt":"2026-07-17T19:55:29","slug":"segment-textual-token","status":"publish","type":"post","link":"https:\/\/www.visual61.com\/site\/segment-textual-token\/","title":{"rendered":"Segment textual (token)"},"content":{"rendered":"

En intel\u00b7lig\u00e8ncia artificial, un token (en catal\u00e0, segment textual) \u00e9s la unitat m\u00ednima en qu\u00e8 un model divideix un text per poder processar-lo. \u00c9s la \u00abmoneda\u00bb amb qu\u00e8 treballen els models de llenguatge.<\/p>\n<\/div>

<\/div>

Qu\u00e8 \u00e9s un token?<\/h2><\/div>

Quan escrius una frase a un model d’IA, aquest no la llegeix tal com ho fem nosaltres. La parteix en fragments anomenats tokens, que poden ser una paraula sencera, un tros de paraula o un signe de puntuaci\u00f3. Despr\u00e9s converteix cada token en n\u00fameros, que \u00e9s l’\u00fanic que una m\u00e0quina sap processar de deb\u00f2.<\/p>\n<\/div>

<\/div>

Per qu\u00e8 s\u00f3n importants?<\/h2><\/div>
    \n
  • El cost de moltes eines d’IA es calcula per tokens consumits<\/li>\n
  • La quantitat de text que un model pot tenir en compte de cop (la finestra de context) es mesura en tokens<\/li>\n
  • Com m\u00e9s tokens, m\u00e9s temps i m\u00e9s recursos de c\u00e0lcul<\/li>\n<\/ul>\n<\/div>
    <\/div>

    Un exemple senzill<\/h2><\/div>

    Una paraula curta i freq\u00fcent com \u00abcasa\u00bb sol ser un sol token. En canvi, una de llarga com \u00abintel\u00b7lig\u00e8ncia\u00bb es pot dividir en diversos. Com a regla aproximada, un token equival a una paraula curta o a una part d’una de llarga, per\u00f2 el nombre exacte dep\u00e8n del model i de l’idioma.<\/p>\n<\/div><\/div><\/div><\/div><\/div>\n","protected":false},"excerpt":{"rendered":"

    En IA, un token (en catal\u00e0, segment textual) \u00e9s la unitat m\u00ednima en qu\u00e8 un model divideix un text per poder processar-lo.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"rank_math_title":"Token o Segment Textual en IA: Definici\u00f3 | Visual61","rank_math_description":"En IA, un token \u00e9s la unitat m\u00ednima en qu\u00e8 un model divideix un text per poder processar-lo. Descobreix com funcionen als models de llenguatge.","rank_math_focus_keyword":"token IA","footnotes":""},"categories":[24],"tags":[37,38,48,53],"class_list":["post-2815","post","type-post","status-publish","format-standard","hentry","category-diccionari","tag-intelligencia-artificial","tag-llm","tag-machine-learning","tag-token"],"_links":{"self":[{"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/posts\/2815","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/comments?post=2815"}],"version-history":[{"count":3,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/posts\/2815\/revisions"}],"predecessor-version":[{"id":3142,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/posts\/2815\/revisions\/3142"}],"wp:attachment":[{"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/media?parent=2815"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/categories?post=2815"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.visual61.com\/site\/wp-json\/wp\/v2\/tags?post=2815"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}