CDCK/MoM cita publicaciones como User1, User2, etc. ¿Problema de calidad del LLM?

Falco · 10 Agosto, 2026 16:49

Gracias por el reporte, este PR maneja este problema:

github.com/discourse/discourse

DEV: Use a realistic few-shot example in the summarizer agent (#42445)

main ← dev-summarizer-realistic-example

merged 06:41PM - 07 Aug 26 UTC

xfalcox

+14 -3

Previously, the summarizer agent's few-shot example used a two-post toy thread w…hose format didn't match the prompts the summarization strategies actually build, and the system prompt described a post format (`N) username message`) that no strategy emits. This change replaces the example with one mirroring a real captured request (discussion title line, posts inside `<input></input>` tags in the real `(N username said: ...` format) paired with a multi-paragraph cited summary demonstrating `[username]({resource_url}/N)` citations, and fixes the post format description to match what the topic, chat, and gist strategies actually send.

El resumen en ese tema ahora se ve mejor.

Tema		Respuestas	Vistas
Moderation tool for formatting code with AI Feature ai	10	399	3 Abril 2025
Code fences don't seem to deal with C comments Support	26	4776	6 Octubre 2015
HTML comments are also summarized by AI Feature ai , ai-summarize	5	286	1 Julio 2026
Posts being made as HTML comments by default? Support	3	667	17 Diciembre 2020
Pasted content gains erroneous line breaks Bug needs-repro	5	252	19 Agosto 2024

CDCK/MoM cita publicaciones como User1, User2, etc. ¿Problema de calidad del LLM?

Temas relacionados