Prompt Condicionado

Create a detailed academic infographic in Spanish that explains how a large language model (LLM) generates text through probabilistic conditioning.

Visual structure (top to bottom):
1. At the top, a wide cloud representing an “Espacio probabilístico latente”, filled with many floating document fragments and text tokens.
2. Inside this space, a neutral abstract AI model icon (not anthropomorphic thinking).
3. A panel labeled:
   “Prompt (condiciones de generación): rol, contexto, audiencia, tarea y restricciones”
   with bullet points for Rol, Contexto, Audiencia, and Tarea.
4. A “Matriz de probabilidades” visualization showing how these conditions reweight token probabilities.
5. A funnel-shaped structure representing progressive restriction of the probability space, with labeled layers:
   Rol → Contexto → Audiencia → Tarea.
6. A label near the funnel:
   “Distribución de probabilidad del siguiente token”.
7. Below the funnel, a label:
   “Texto generado token por token”.
8. At the bottom, a reduced area labeled:
   “Espacio de probabilidad reducido”,
   containing an example prompt about explaining what a gene is in molecular genetics for undergraduate genomics students.
9. Final label at the bottom:
   “Salida textual condicionada”.

Style:
- Clean academic infographic
- Soft blue and neutral tones
- Clear arrows indicating flow
- No equations
- No anthropomorphic reasoning metaphors
- Educational, university-level design