{"translation-revision-date":"2026-08-26T18:40:05+00:00","generator":"make-json.php/1.0","source":"modules/content/content.js","domain":"messages","locale_data":{"messages":{"":{"domain":"messages","lang":"es_ES","plural-forms":"nplurals=2; plural=(n != 1);"},"Clear the title list?":["\u00bfVaciar la lista de t\u00edtulos?"],"The whole list will be cleared. Articles already added to the queue will not be affected.":["Se vaciar\u00e1 toda la lista. Los art\u00edculos ya a\u00f1adidos a la cola no se ver\u00e1n afectados."],"Yes, clear":["S\u00ed, vaciar"],"No titles":["Sin t\u00edtulos"],"Add at least one title to be able to generate.":["A\u00f1ade al menos un t\u00edtulo para poder generar."],"Adding\u2026":["A\u00f1adiendo\u2026"],"Add to the generation queue":["A\u00f1adir a la cola de generaci\u00f3n"],"Checking coverage\u2026":["Comprobando cobertura\u2026"],"Generating article %1$s of %2$s\u2026":["Generando art\u00edculo %1$s de %2$s\u2026"],"Connection error. Press Generate again.":["Error de conexi\u00f3n. Pulsa Generar de nuevo."],"Generation completed":["Generaci\u00f3n completada"],"Updating queue\u2026":["Actualizando cola\u2026"],"Completed":["Completado"],"Reload the page to see the new articles.":["Recarga la p\u00e1gina para ver los nuevos art\u00edculos."],"Cancel pending articles":["Cancelar art\u00edculos pendientes"],"All articles still queued and not generated will be cancelled. Those already finished will not be affected.":["Se cancelar\u00e1n todos los art\u00edculos a\u00fan en cola y sin generar. Los ya finalizados no se ver\u00e1n afectados."],"Generate articles":["Generar art\u00edculos"],"Finished: %s":["Finalizado: %s"],"Generated %s":["Generado %s"],"Failed: %s":["Fall\u00f3: %s"],"Failed %s":["Fall\u00f3 %s"],"Failed":["Fallidos"],"Generating\u2026":["Generando\u2026"],"Queued":["En cola"],"No title":["Sin t\u00edtulo"],"Write at least one title in the titles field to preview the prompt.":["Escribe al menos un t\u00edtulo en el campo de t\u00edtulos para previsualizar el prompt."],"See the exact prompt that will be sent to DALL\u00b7E":["Ve el prompt exacto que se enviar\u00e1 a DALL\u00b7E"],"Error":["Error"],"The preview could not be generated.":["No se pudo generar la vista previa."],"Includes article summary":["Incluye resumen del art\u00edculo"],"The summary will be added automatically when generating":["El resumen se a\u00f1adir\u00e1 autom\u00e1ticamente al generar"],"Title:":["T\u00edtulo:"],"Style:":["Estilo:"],"Characters:":["Caracteres:"],"Full prompt sent to %s:":["Prompt completo enviado a %s:"],"Copy prompt":["Copiar prompt"],"Image prompt \u2014 preview":["Prompt de imagen \u2014 vista previa"],"Connection error":["Error de conexi\u00f3n"],"Could not connect to the server.":["No se pudo conectar con el servidor."],"Copied!":["\u00a1Copiado!"],"Download all prompts in a single documented .txt file":["Descarga todos los prompts en un \u00fanico archivo .txt documentado"],"No data":["Sin datos"],"The saved prompt was not found.":["No se encontr\u00f3 el prompt guardado."],"fallback":["fallback"],"main":["principal"],"Model:":["Modelo:"],"Size:":["Tama\u00f1o:"],"Quality:":["Calidad:"],"Yes":["S\u00ed"],"No":["No"],"Error:":["Error:"],"gpt-image-1 connection error:":["Error de conexi\u00f3n gpt-image-1:"],"Main keyword":["Keyword principal"],"Entity / topic":["Entidad / tema"],"Geography":["Geograf\u00eda"],"Intent":["Intenci\u00f3n"],"Structure (H2s)":["Estructura (H2s)"],"Intro":["Intro"],"Template used":["Plantilla usada"],"Tweaks for this article":["Matices de este art\u00edculo"],"Context used to build the scene:":["Contexto usado para construir la escena:"],"Fallback activated":["Fallback activado"],"No fallback":["Sin fallback"],"Prompt sent to %s:":["Prompt enviado a %s:"],"How the model interpreted it (revised prompt):":["C\u00f3mo lo interpret\u00f3 el modelo (prompt revisado):"],"Copy revised prompt":["Copiar prompt revisado"],"Image prompt \u2014 sent with the article":["Prompt de imagen \u2014 enviado con el art\u00edculo"],"Architecture:":["Arquitectura:"],"By sections (v4.4)":["Por secciones (v4.4)"],"Single call (legacy)":["Llamada \u00fanica (legacy)"],"Temperature:":["Temperatura:"],"Result:":["Resultado:"],"2 passes":["2 pasadas"],"System prompt (%s characters):":["System prompt (%s caracteres):"],"Copy system":["Copiar system"],"1. Outline (article structure)":["1. Outline (estructura del art\u00edculo)"],"Outline JSON response:":["Respuesta JSON del outline:"],"2. Introduction":["2. Introducci\u00f3n"],"%1$s. Section %2$s: \u00ab%3$s\u00bb (%4$s words)":["%1$s. Secci\u00f3n %2$s: \u00ab%3$s\u00bb (%4$s palabras)"],"%s. Closing / conclusion":["%s. Cierre / conclusi\u00f3n"],"User prompt (%s characters):":["User prompt (%s caracteres):"],"Copy user":["Copiar user"],"2nd pass prompt (expand \u00b7 %s characters":["Prompt de la 2\u00aa pasada (expand \u00b7 %s caracteres"],"temp %s":["temp %s"],"max %s tokens":["m\u00e1x %s tokens"],"Copy expand":["Copiar expand"],"Text prompt \u2014 sent with the article":["Prompt de texto \u2014 enviado con el art\u00edculo"],"No saved prompt was found for this article.":["No se encontr\u00f3 ning\u00fan prompt guardado para este art\u00edculo."],"No text prompt was saved for this article.":["No se guard\u00f3 prompt de texto para este art\u00edculo."],"No image was generated for this article (or the prompt was not saved).":["No se gener\u00f3 imagen para este art\u00edculo (o el prompt no se guard\u00f3)."],"Text prompts":["Prompts de texto"],"Image prompt":["Prompt de imagen"],"Prompts \u2014 article #%s":["Prompts \u2014 art\u00edculo #%s"]," (%s characters):":[" (%s caracteres):"],"There is no payload available to download.":["No hay payload disponible para descargar."],"END OF TEXT DUMP \u2014 IMAGE DUMP FOLLOWS":["FIN DEL DUMP DE TEXTO \u2014 A CONTINUACI\u00d3N, DUMP DE IMAGEN"],"There is no text or image data to download.":["No hay datos de texto ni de imagen para descargar."],"Unknown dump type: %s":["Tipo de dump desconocido: %s"],"Downloaded!":["\u00a1Descargado!"],"The file could not be generated: %s":["No se pudo generar el archivo: %s"],"Characters: %s":["Caracteres: %s"],"Latency: %s":["Latencia: %s"],"(empty)":["(vac\u00edo)"],"Payload is not of type TEXT (kind=%s)":["El payload no es de tipo TEXTO (kind=%s)"],"4LINKING \u2014 TEXT PROMPTS DUMP":["4LINKING \u2014 DUMP DE PROMPTS DE TEXTO"],"Plugin version:":["Versi\u00f3n del plugin:"],"Created:":["Creado:"],"Modified:":["Modificado:"],"Post status:":["Estado del post:"],"Passes:":["Pasadas:"],"2 (with expand)":["2 (con expand)"],"Failed sections:":["Secciones fallidas:"],"Length warning:":["Aviso de longitud:"],"YES (article did not reach the target)":["S\u00cd (el art\u00edculo no alcanz\u00f3 el objetivo)"],"Generation start:":["Inicio de generaci\u00f3n:"],"Generation end:":["Fin de generaci\u00f3n:"],"Total duration:":["Duraci\u00f3n total:"],"OpenAI calls:":["Llamadas a OpenAI:"],"Total tokens:":["Tokens totales:"],"API latency:":["Latencia de la API:"],"sum of all calls":["suma de todas las llamadas"],"KB coverage:":["Cobertura KB:"],"entries used (high/medium confidence)":["fichas usadas (confianza alta/media)"],"%s weak signal":["%s se\u00f1al d\u00e9bil"],"KB selection:":["Selecci\u00f3n KB:"],"entries injected into the prompt":["fichas inyectadas al prompt"],"%s below threshold":["%s bajo umbral"],"Claims without KB support:":["Afirmaciones sin respaldo KB:"],"review":["revisar"],"Editorial audit:":["Auditor\u00eda editorial:"],"Downloaded:":["Descargado:"],"JOB PARAMETERS":["PAR\u00c1METROS DEL JOB"],"Template:":["Plantilla:"],"(no name)":["(sin nombre)"],"(no template \u2014 free form)":["(sin plantilla \u2014 formulario libre)"],"Length:":["Longitud:"],"Tone:":["Tono:"],"Structure:":["Estructura:"],"Audience:":["Audiencia:"],"Point of view:":["Punto de vista:"],"Reading level:":["Nivel de lectura:"],"Language:":["Idioma:"],"Main keyword:":["Keyword principal:"],"Search intent:":["Intenci\u00f3n de b\u00fasqueda:"],"Secondary keywords:":["Keywords secundarias:"],"Forced H2:":["H2 forzados:"],"(yes \u2014 see configuration in the form)":["(s\u00ed \u2014 ver configuraci\u00f3n en el formulario)"],"Words to avoid:":["Palabras a evitar:"],"Assigned KB:":["KB asignada:"],"LOW COVERAGE on this title (score %s)":["COBERTURA BAJA en este t\u00edtulo (score %s)"],"(none)":["(ninguna)"],"Image model:":["Modelo de imagen:"],"This file contains ALL the prompts sent to OpenAI during":["Este archivo contiene TODOS los prompts enviados a OpenAI durante"],"the generation of this article. It is meant for auditing:":["la generaci\u00f3n de este art\u00edculo. Est\u00e1 pensado para auditor\u00eda:"],"it lets you review what was sent to the model and diagnose problems":["permite revisar qu\u00e9 se mand\u00f3 al modelo y diagnosticar problemas"],"of fidelity, style or structure without having to rebuild anything.":["de fidelidad, estilo o estructura sin tener que reconstruir nada."],"The blocks are presented in the ACTUAL EXECUTION ORDER:":["Los bloques se presentan en el ORDEN REAL DE EJECUCI\u00d3N:"],"1. System prompt (shared by all calls).":["1. System prompt (compartido por todas las llamadas)."],"2. Outline \u2192 generates the article JSON structure.":["2. Outline \u2192 genera la estructura JSON del art\u00edculo."],"3. Intro \u2192 first paragraph of the article.":["3. Intro \u2192 primer p\u00e1rrafo del art\u00edculo."],"4..N. One call per H2 section.":["4..N. Una llamada por cada secci\u00f3n H2."],"N+1. Closing / conclusion.":["N+1. Cierre / conclusi\u00f3n."],"Lines starting with `#` are annotations; the rest is":["Las l\u00edneas que empiezan por `#` son anotaciones; el resto es"],"literal text sent to the model.":["texto literal enviado al modelo."],"Characters: %s  \u00b7  Role: system  \u00b7  Shared across all calls":["Caracteres: %s  \u00b7  Rol: system  \u00b7  Compartido en todas las llamadas"],"The system prompt is sent with EVERY pipeline call (outline,\nintro, each section, closing). It defines the model persona and the\nstyle rules (the 19 anti-AI rules). It does not contain the KB.":["El system prompt se env\u00eda con CADA llamada del pipeline (outline,\nintro, cada secci\u00f3n, cierre). Define la persona del modelo y las\nreglas de estilo (las 19 reglas anti-IA). No contiene la KB."],"Role: user  \u00b7  Pipeline call #1":["Rol: user  \u00b7  Llamada n\u00ba 1 del pipeline"],"This call generates the STRUCTURE of the article (JSON with the\nlist of H2 sections, their key_points and which required elements\neach one carries: table, FAQ, checklist, etc.).\nIMPORTANT: by default this prompt does NOT receive the injected KB.\nThe model decides the structure without knowing what the knowledge\nbase contains, which conditions the rest of the article.":["Esta llamada genera la ESTRUCTURA del art\u00edculo (JSON con la\nlista de secciones H2, sus key_points y qu\u00e9 elementos\nrequeridos lleva cada una: tabla, FAQ, checklist, etc.).\nIMPORTANTE: por defecto este prompt NO recibe la KB inyectada.\nEl modelo decide la estructura sin saber qu\u00e9 contiene la base\nde conocimiento, lo que condiciona todo el resto del art\u00edculo."],"OUTLINE \u2014 MODEL RESPONSE (JSON)":["OUTLINE \u2014 RESPUESTA DEL MODELO (JSON)"],"Characters: %s  \u00b7  Role: assistant  \u00b7  Output of call #1":["Caracteres: %s  \u00b7  Rol: assistant  \u00b7  Output de la llamada n\u00ba 1"],"JSON returned by the model with the final article structure.\nEach section here will trigger a separate call in the\nfollowing blocks.":["JSON devuelto por el modelo con la estructura definitiva del\nart\u00edculo. Cada secci\u00f3n de aqu\u00ed desencadenar\u00e1 una llamada\nseparada en los bloques siguientes."],"INTRODUCTION \u2014 USER PROMPT":["INTRODUCCI\u00d3N \u2014 USER PROMPT"],"Role: user  \u00b7  Generates the paragraph before the first H2":["Rol: user  \u00b7  Genera el p\u00e1rrafo previo al primer H2"],"Call that produces the introductory paragraph of the article\n(before the first H2). It is usually short: 100-180 words.":["Llamada que produce el p\u00e1rrafo introductorio del art\u00edculo\n(antes del primer H2). Suele ser corto: 100-180 palabras."],"INTRODUCTION \u2014 MODEL RESPONSE":["INTRODUCCI\u00d3N \u2014 RESPUESTA DEL MODELO"],"Characters: %s  \u00b7  Role: assistant":["Caracteres: %s  \u00b7  Rol: assistant"],"HTML of the introduction as returned by the model,\nbefore the filtering and cleanup pass. Useful to tell apart\nwhat OpenAI produced from what apply_avoid_filter rewrote.":["HTML de la introducci\u00f3n tal como lo devolvi\u00f3 el modelo,\nantes del paso de filtros y limpieza. \u00datil para distinguir\nlo que produjo OpenAI de lo que reescribi\u00f3 apply_avoid_filter."],"Position: %1$s  \u00b7  Target words: %2$s":["Posici\u00f3n: %1$s  \u00b7  Palabras objetivo: %2$s"],"Section \u00ab%s\u00bb.":["Secci\u00f3n \u00ab%s\u00bb."],"One independent call per H2 of the article.\nThis is where the KB (verified information) is injected: each\nsection receives the KB block with the entries relevant\nto its topic, along with the mandatory fidelity rules.":["Una llamada independiente por cada H2 del art\u00edculo.\nEs donde se inyecta la KB (informaci\u00f3n verificada): cada\nsecci\u00f3n recibe el bloque KB con las fichas relevantes\na su tema, junto con las reglas obligatorias de fidelidad."],"SECTION %s \u2014 USER PROMPT":["SECCI\u00d3N %s \u2014 USER PROMPT"],"SECTION %s \u2014 MODEL RESPONSE":["SECCI\u00d3N %s \u2014 RESPUESTA DEL MODELO"],"Characters: %1$s  \u00b7  Role: assistant  \u00b7  H2: \u00ab%2$s\u00bb":["Caracteres: %1$s  \u00b7  Rol: assistant  \u00b7  H2: \u00ab%2$s\u00bb"],"HTML returned by the model for this section, before\nassembly and filters. Check here whether a quality\ncomplaint comes from the prompt-builder or the model.":["HTML que devolvi\u00f3 el modelo para esta secci\u00f3n, antes\ndel ensamblado y los filtros. Verificar aqu\u00ed si una\nqueja de calidad viene del prompt-builder o del modelo."],"CLOSING / CONCLUSION \u2014 USER PROMPT":["CIERRE / CONCLUSI\u00d3N \u2014 USER PROMPT"],"Role: user  \u00b7  Last call in the pipeline":["Rol: user  \u00b7  \u00daltima llamada del pipeline"],"Generates the closing paragraph of the article. It usually does not\nreceive the injected KB because its function is to synthesize, not to add\nnew data.":["Genera el p\u00e1rrafo de cierre del art\u00edculo. No suele recibir\nla KB inyectada porque su funci\u00f3n es sintetizar, no aportar\ndatos nuevos."],"CLOSING / CONCLUSION \u2014 MODEL RESPONSE":["CIERRE / CONCLUSI\u00d3N \u2014 RESPUESTA DEL MODELO"],"Closing as returned by the model. The conclusion is\nthe point most exposed to clich\u00e9s (\"next level\",\n\"makes the difference\"); comparing it with the final HTML\nshows what the filter rewrote.":["Cierre tal como lo devolvi\u00f3 el modelo. La conclusi\u00f3n es\nel punto m\u00e1s expuesto a clich\u00e9s (\"siguiente nivel\",\n\"marcar\u00e1 la diferencia\"); compararla con el HTML final\npermite ver qu\u00e9 reescribi\u00f3 el filtro."],"USER PROMPT (LEGACY \u00b7 SINGLE CALL)":["USER PROMPT (LEGACY \u00b7 LLAMADA \u00daNICA)"],"Characters: %s  \u00b7  Role: user  \u00b7  Legacy architecture (single call)":["Caracteres: %s  \u00b7  Rol: user  \u00b7  Arquitectura legacy (una sola llamada)"],"Legacy architecture: the whole article is generated in a\nsingle call. The separate sections of the v4.4 pipeline\ndo not apply here.":["Arquitectura legacy: todo el art\u00edculo se genera en una\n\u00fanica llamada. Las secciones por separado del pipeline v4.4\nno aplican aqu\u00ed."],"2ND PASS PROMPT (EXPAND)":["PROMPT DE 2\u00aa PASADA (EXPAND)"],"Temp: %s":["Temp: %s"],"Max tokens: %s":["M\u00e1x tokens: %s"],"Second pass used to extend articles that did not\nreach the target length in the first call.":["Segunda pasada usada para extender art\u00edculos que no\nalcanzaron la longitud objetivo en la primera llamada."],"Entries in the base (total):":["Fichas en la base (total):"],"Above threshold:":["Por encima del umbral:"],"Below threshold:":["Por debajo del umbral:"],"Minimum threshold applied:":["Umbral m\u00ednimo aplicado:"],"Configured limit (top-N):":["L\u00edmite configurado (top-N):"],"Entries that entered the prompt:":["Fichas que entraron al prompt:"],"INJECTED ENTRIES (ordered by similarity)":["FICHAS INYECTADAS (ordenadas por similitud)"],"(no title)":["(sin t\u00edtulo)"],"FIRST ENTRY EXCLUDED BY THE CUT":["PRIMERA FICHA EXCLUIDA POR EL CORTE"],"(It was just after the top-N. If its score is close\n   to the last included one, the cut was marginal.)":["(Qued\u00f3 justo despu\u00e9s del top-N. Si su score est\u00e1 cerca\n   del \u00faltimo incluido, el corte fue marginal.)"],"TOP 3 BELOW THRESHOLD (discarded for low similarity)":["TOP 3 POR DEBAJO DEL UMBRAL (descartadas por baja similitud)"],"(These did not enter the prompt because their score was\n   below the minimum threshold. If you think they should\n   be there, consider lowering the threshold or rewriting the\n   entry title to match the article one.)":["(Estas no entraron al prompt porque su score qued\u00f3\n   por debajo del umbral m\u00ednimo. Si crees que deb\u00edan\n   estar, considera bajar el umbral o reescribir el\n   t\u00edtulo de la ficha para que case con el del art\u00edculo.)"],"PRE-GEN AUDIT \u2014 KB ENTRY SELECTION":["AUDITOR\u00cdA PRE-GEN \u2014 SELECCI\u00d3N DE FICHAS KB"],"Which entries were chosen, with what similarity, what was left out":["Qu\u00e9 fichas se eligieron, con qu\u00e9 similitud, qu\u00e9 qued\u00f3 fuera"],"Diagnostic made BEFORE generating (during prompt\nbuilding): the semantic engine measures the similarity between the\narticle title and each entry of the base, sorts by\nscore and keeps the first ones up to the limit. This\nblock reports which entries passed the cut, with what scores,\nwhich one came right after and what was discarded by threshold.\nIf the article does not use an entry you think it should,\nlook here first: if the entry is included, the problem\nis the model (it did not use it); if it is not, the problem is the\nselection (it did not reach the prompt).":["Diagn\u00f3stico hecho ANTES de generar (durante la construcci\u00f3n\ndel prompt): el motor sem\u00e1ntico mide la similitud entre el\nt\u00edtulo del art\u00edculo y cada ficha de la base, ordena por\nscore y se queda con las primeras hasta el l\u00edmite. Este\nbloque cuenta qu\u00e9 fichas pasaron el corte, con qu\u00e9 scores,\ncu\u00e1l qued\u00f3 justo despu\u00e9s y qu\u00e9 se descart\u00f3 por umbral.\nSi el art\u00edculo no usa una ficha que crees que deb\u00eda usar,\nmira aqu\u00ed primero: si la ficha est\u00e1 incluida, el problema\nes del modelo (no la us\u00f3); si no est\u00e1, el problema es de\nla selecci\u00f3n (no lleg\u00f3 al prompt)."],"Entries fed to the outline:":["Fichas alimentadas al outline:"],"Entries reflected in the output:":["Fichas reflejadas en el output:"],"Unused entries:":["Fichas no utilizadas:"],"Coverage rate:":["Tasa de cobertura:"],"DETAIL PER ENTRY (how usage was detected)":["DETALLE POR FICHA (c\u00f3mo se detect\u00f3 el uso)"],"(Three tests in order, in decreasing CONFIDENCE:\n  \u00b7 HIGH  \u2014 the full title appears verbatim.\n  \u00b7 MEDIUM \u2014 a 3-word sequence of the title appears.\n  \u00b7 LOW  \u2014 only one distinctive token of the title (\u22657 chars and infrequent in the KB).\n Only HIGH+MEDIUM count as \"used\" \u2014 LOW goes to its own bucket\n because it has too many false positives to be reliable.)":["(Tres tests en orden, en CONFIANZA decreciente:\n  \u00b7 ALTA  \u2014 el t\u00edtulo completo aparece verbatim.\n  \u00b7 MEDIA \u2014 una secuencia de 3 palabras del t\u00edtulo aparece.\n  \u00b7 BAJA  \u2014 solo un token distintivo del t\u00edtulo (\u22657 chars y poco frecuente en la KB).\n Solo ALTA+MEDIA cuentan como \"usadas\" \u2014 BAJA va a su propio bucket\n porque tiene demasiados falsos positivos para ser fiable.)"],"verbatim":["verbatim"],"3-token phrase":["frase 3-token"],"single token":["token \u00fanico"],"USED \u2014 high or medium confidence (%s):":["USADAS \u2014 confianza alta o media (%s):"],"WEAK SIGNAL \u2014 only one distinctive token appears (%s):":["SE\u00d1AL D\u00c9BIL \u2014 solo un token distintivo aparece (%s):"],"(The entry may have been used, or it may be coincidence.\n   Check whether the entry central topic is in the article.)":["(Puede que la ficha se usara, puede que sea coincidencia.\n   Revisar si el tema central de la ficha est\u00e1 en el art\u00edculo.)"],"NOT USED \u2014 no test matches (%s):":["NO USADAS \u2014 ning\u00fan test casa (%s):"],"Entries that were sent but did not appear in the article:":["Fichas que se enviaron pero no aparecieron en el art\u00edculo:"],"(All fed entries have a trace in the article.)":["(Todas las fichas alimentadas tienen rastro en el art\u00edculo.)"],"POST-GEN AUDIT \u2014 KB COVERAGE (approximate)":["AUDITOR\u00cdA POST-GEN \u2014 COBERTURA KB (aproximada)"],"Analysis of which entries fed the model and which appeared":["An\u00e1lisis de qu\u00e9 fichas alimentaron al modelo y cu\u00e1les aparecieron"],"Not a prompt. It is an analysis made after generating the\narticle: it checks which KB titles were sent to the outline and\nwhich left a trace in the final HTML. Detection is\nAPPROXIMATE \u2014 it looks for distinctive words of the title within\nthe text. There can be false positives (a generic term\nshared among several entries) and false negatives (an\nentry whose concept was reflected but heavily paraphrased).\nUseful as an order of magnitude and an iteration hint, not as a\nprecise metric.":["No es un prompt. Es un an\u00e1lisis hecho despu\u00e9s de generar el\nart\u00edculo: comprueba qu\u00e9 t\u00edtulos KB se enviaron al outline y\ncu\u00e1les dejaron rastro en el HTML final. La detecci\u00f3n es\nAPROXIMADA \u2014 busca palabras distintivas del t\u00edtulo dentro\ndel texto. Puede haber falsos positivos (un t\u00e9rmino gen\u00e9rico\ncompartido entre varias fichas) y falsos negativos (una\nficha cuyo concepto se reflej\u00f3 pero parafraseado fuerte).\n\u00datil como orden de magnitud y pista de iteraci\u00f3n, no como\nm\u00e9trica precisa."],"Claims detected in the article whose concrete value\nDOES NOT appear in the KB block injected to the model.":["Afirmaciones detectadas en el art\u00edculo cuyo valor concreto\nNO aparece en el bloque KB inyectado al modelo."],"Unsupported percentages":["Porcentajes sin respaldo"],"Unsupported years / dates":["A\u00f1os / fechas sin respaldo"],"Unsupported absolute figures":["Cifras absolutas sin respaldo"],"Unsupported authority attributions":["Atribuciones de autoridad sin respaldo"],"Claim:":["Afirmaci\u00f3n:"],"Context:":["Contexto:"],"POST-GEN AUDIT \u2014 CLAIMS WITHOUT KB SUPPORT":["AUDITOR\u00cdA POST-GEN \u2014 AFIRMACIONES SIN RESPALDO EN KB"],"Possible quantitative or authority fabrications \u2014 review before publishing":["Posibles invenciones cuantitativas o de autoridad \u2014 revisar antes de publicar"],"Heuristic detector for rule 19 of the system prompt: the AI\nmust not invent percentages, dates, studies or\nauthorities; it can only use those in the KB or in\nthe advanced instructions. This block scans the final HTML\nlooking for that kind of claim and checks whether the\nconcrete value (the exact number, the year, the name)\nappears in the KB block injected to the model. If\nit does not appear, the claim is reported as a candidate for\nfabrication.\nIMPORTANT: it is heuristic. There are false positives (a legitimate\ngeneric number, e.g. \"the 6 modules\" when it is not\nliteral in the KB but is implicit) and possible false\nnegatives (an invented number that happens to coincide\nwith a legitimate one in the KB). Use it as a review list,\nnot as a verdict.":["Detector heur\u00edstico de la regla 19 del system prompt: la IA\nno debe inventar porcentajes, fechas, estudios ni\nautoridades; solo puede usar las que est\u00e9n en la KB o en\nlas instrucciones avanzadas. Este bloque escanea el HTML\nfinal buscando ese tipo de afirmaciones y comprueba si el\nvalor concreto (el n\u00famero exacto, el a\u00f1o, el nombre)\naparece en el bloque KB que se le inyect\u00f3 al modelo. Si\nno aparece, la afirmaci\u00f3n se reporta como candidata a\ninvenci\u00f3n.\nIMPORTANTE: es heur\u00edstico. Hay falsos positivos (un n\u00famero\ngen\u00e9rico leg\u00edtimo, p. ej. \"los 6 m\u00f3dulos\" cuando no est\u00e1\nliteral en la KB pero est\u00e1 impl\u00edcito) y posibles falsos\nnegativos (un n\u00famero inventado que casualmente coincide\ncon uno leg\u00edtimo en la KB). \u00dasalo como lista de revisi\u00f3n,\nno como veredicto."],"(The per-family detail is in \"Detected patterns\".)":["(El detalle por familia est\u00e1 en \"Patrones detectados\".)"],"APPLIED REWRITES: none.":["REESCRITURAS APLICADAS: ninguna."],"DETECTED PATTERNS":["PATRONES DETECTADOS"],"(Families of generic SEO patterns detected in the text.\n Some were rewritten automatically, others feed\n the editorial audit panel as alerts.)":["(Familias de patrones SEO gen\u00e9ricos detectadas en el texto.\n Algunas se reescribieron autom\u00e1ticamente, otras alimentan\n al panel de auditor\u00eda editorial como alertas.)"],"DETECTED PATTERNS: none.":["PATRONES DETECTADOS: ninguno."],"BLACKLIST WORDS THE MODEL USED":["PALABRAS DE LA LISTA NEGRA QUE EL MODELO US\u00d3"],"TEXT SHAPE":["FORMA DEL TEXTO"],"Total sentences:":["Frases totales:"],"Short sentences (<13w):":["Frases cortas (<13p):"],"Median words/sentence:":["Mediana palabras/frase:"],"H2 openers (total):":["Aperturas H2 (total):"],"Cloned H2 openers:":["Aperturas H2 cl\u00f3nicas:"],"Authorial voice moves:":["Movimientos de voz autorial:"],"EDITORIAL AUDIT BY CATEGORY":["AUDITOR\u00cdA EDITORIAL POR CATEGOR\u00cdA"],"Overall risk:":["Riesgo global:"],"(no label)":["(sin etiqueta)"],"low":["bajo"],"Detectors that fired:":["Detectores que dispararon:"],"Concrete phrases detected:":["Frases concretas detectadas:"],"(%s more, truncated)":["(%s m\u00e1s, recortados)"],"Auditor recommendations:":["Recomendaciones del auditor:"],"POST-GEN AUDIT \u2014 FILTERS AND EDITORIAL QUALITY":["AUDITOR\u00cdA POST-GEN \u2014 FILTROS Y CALIDAD EDITORIAL"],"Applied rewrites + passive detectors + editorial audit":["Reescrituras aplicadas + detectores pasivos + auditor\u00eda editorial"],"This block reports what happened AFTER the model\nreturned the text: the deterministic filter (apply_avoid_filter)\nrewrites some patterns and leaves others as alerts. Combined\nwith the MODEL RESPONSES of the previous blocks, it lets you\ntell apart three kinds of problem:\n  1. the model did not generate something well (look at the raw response),\n  2. the filter should have rewritten it and did not (look at\n     \"Detected patterns\"),\n  3. the filter rewrote it but the result is worse (compare\n     the raw response with the final HTML of the post).":["Este bloque cuenta lo que pas\u00f3 DESPU\u00c9S de que el modelo\ndevolviera el texto: el filtro determinista (apply_avoid_filter)\nreescribe algunos patrones y deja otros como alertas. Combinado\ncon las RESPUESTAS DEL MODELO de los bloques anteriores, permite\ndistinguir tres clases de problema:\n  1. el modelo no gener\u00f3 algo bien (mirar respuesta cruda),\n  2. el filtro deber\u00eda haberlo reescrito y no lo hizo (mirar\n     \"Patrones detectados\"),\n  3. el filtro lo reescribi\u00f3 pero el resultado es peor (comparar\n     respuesta cruda con HTML final del post)."],"Introduction":["Introducci\u00f3n"],"Section %s":["Secci\u00f3n %s"],"Conclusion":["Conclusi\u00f3n"],"Call":["Llamada"],"Latency":["Latencia"],"Model returned by the API:":["Modelo devuelto por la API:"],"OPENAI CALL METRICS":["M\u00c9TRICAS DE LLAMADAS A OPENAI"],"Latency and tokens per pipeline block":["Latencia y tokens por bloque del pipeline"],"Pt = prompt tokens, Ct = completion tokens, Tt = total tokens.\nFinish: \"stop\" is normal. \"length\" indicates the response\nwas truncated by max_tokens \u2014 you would need to raise that limit. Other\nrare values can be a flag to investigate.":["Pt = prompt tokens, Ct = completion tokens, Tt = total tokens.\nFinish: \"stop\" es lo normal. \"length\" indica que la respuesta\nse trunc\u00f3 por max_tokens \u2014 habr\u00eda que subir ese l\u00edmite. Otros\nvalores raros pueden ser bandera para investigar."],"Payload is not of type IMAGE (kind=%s)":["El payload no es de tipo IMAGEN (kind=%s)"],"4LINKING \u2014 IMAGE PROMPTS DUMP":["4LINKING \u2014 DUMP DE PROMPTS DE IMAGEN"],"Fallback:":["Fallback:"],"YES (main model failed)":["S\u00cd (el modelo principal fall\u00f3)"],"no":["no"],"Generated:":["Generado:"],"This file contains the prompts sent to OpenAI to":["Este archivo contiene los prompts enviados a OpenAI para"],"generate the featured image of this article, along with the":["generar la imagen destacada de este art\u00edculo, junto con el"],"scene context used to build them.":["contexto de escena que se us\u00f3 para construirlos."],"Entity / topic:":["Entidad / tema:"],"Geography:":["Geograf\u00eda:"],"Intent:":["Intenci\u00f3n:"],"Structure (H2s):":["Estructura (H2s):"],"Intro:":["Intro:"],"Template instructions:":["Instrucciones de la plantilla:"],"Article tweaks:":["Matices del art\u00edculo:"],"SCENE CONTEXT":["CONTEXTO DE ESCENA"],"Data used to build the visual prompt":["Datos usados para construir el prompt visual"],"Context extracted from the article and the template, fed to the\nscene builder to produce the prompt sent to the\nimage model. It is not sent literally to OpenAI: it is the\nraw material from which the prompt is built.":["Contexto extra\u00eddo del art\u00edculo y la plantilla, alimentado al\nconstructor de escena para producir el prompt enviado al\nmodelo de imagen. No se env\u00eda literalmente a OpenAI: es la\nmateria prima a partir de la cual se construye el prompt."],"PROMPT SENT TO THE IMAGE MODEL":["PROMPT ENVIADO AL MODELO DE IMAGEN"],"Characters: %1$s  \u00b7  Model: %2$s":["Caracteres: %1$s  \u00b7  Modelo: %2$s"],"Literal text sent to the image generation endpoint.":["Texto literal enviado al endpoint de generaci\u00f3n de imagen."],"PROMPT REWRITTEN BY THE MODEL (revised_prompt)":["PROMPT REESCRITO POR EL MODELO (revised_prompt)"],"Characters: %s  \u00b7  Only appears with dall-e-3":["Caracteres: %s  \u00b7  Solo aparece con dall-e-3"],"When the model is dall-e-3, OpenAI internally rewrites the\nprompt before generating the image. This is that rewritten text,\nuseful to understand what the model interpreted.":["Cuando el modelo es dall-e-3, OpenAI reescribe internamente el\nprompt antes de generar la imagen. Este es ese texto reescrito,\n\u00fatil para entender qu\u00e9 interpret\u00f3 el modelo."],"DEBUG \u2014 RAW REQUEST":["DEBUG \u2014 REQUEST CRUDA"],"Full body sent to the OpenAI endpoint":["Body completo enviado al endpoint de OpenAI"],"Exact JSON sent in the HTTP call. Useful to see\nparameters like size, quality, response_format, etc.":["JSON exacto que se envi\u00f3 en la llamada HTTP. \u00datil para ver\npar\u00e1metros como size, quality, response_format, etc."],"DEBUG \u2014 gpt-image-1 RESPONSE":["DEBUG \u2014 RESPUESTA DE gpt-image-1"],"Technical information of the call to the main model":["Informaci\u00f3n t\u00e9cnica de la llamada al modelo principal"],"Raw JSON with metadata of the call to the main model\n(http_code, size, quality, errors). Useful when there is\nfallback to the backup model.":["JSON crudo con metadatos de la llamada al modelo principal\n(http_code, tama\u00f1o, calidad, errores). \u00datil cuando hay\nfallback al modelo de respaldo."],"DEBUG \u2014 gpt-image-1 CONNECTION ERROR":["DEBUG \u2014 ERROR DE CONEXI\u00d3N gpt-image-1"],"Error captured when trying to use the main model":["Error capturado al intentar usar el modelo principal"],"If present, the main model failed completely and\nthe automatic fallback was used.":["Si est\u00e1 presente, el modelo principal fall\u00f3 por completo y\nse us\u00f3 el fallback autom\u00e1tico."],"Category and tags apply only to the \"Post\" type.":["Categor\u00eda y etiquetas se aplican solo al tipo \"Entrada\"."],"For \"%s\" you can assign taxonomies from the editor after generating the article.":["Para \"%s\" puedes asignar las taxonom\u00edas desde el editor tras generar el art\u00edculo."]}}}