how many screenshots are used by WorldModel to plan next instruction?

[XingWang1234]

hello,

seems that only the latest screenshot by WorldModelt to plan next instruction?

Thanks.

[XingWang1234]

in WORLD_MODEL_GENERAL_EXAMPLES, for vision, some value is SCREENSHOTS, some value is SCREENSHOT,
what are the difference?

Current state:
external_observations:
  vision: '[SCREENSHOTS]'
internal_state:
  agent_outputs: []
  user_inputs: []

Current state:
external_observations:
  vision: '[SCREENSHOT]'
internal_state:
  agent_outputs: []
  user_inputs: []