In a single illustrative episode (seed 2), mean EFE after sticker removal was 12.00 lower than before removal after 500k training steps

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Mean EFE before sticker removal across 80 evaluations: 79.33 ± 4.34finding0.881
  Baseline EFE when sticker is present, used for comparison
• EFE decrease after sticker removal is statistically significant (Wilcoxon p = 6.33×10⁻⁹) across 80 evaluationsfinding0.860
  Confirms that EFE systematically decreases after sticker removal, validating the self-prior as internal criterion
• Untrained model (0 training steps) shows no clear EFE difference before and after sticker removal (Δ = +1.70)finding0.856
  Control showing that the EFE signal is learned, not inherent to the architecture
• Sticker-removal success rate stayed near 20% in the early phase of trainingfinding0.812
  Shows learning progression from chance-level to functional behavior
• Mean hand-sticker distance decreased gradually across 500k training steps, including before removal probability exceeded 50%finding0.809
  Suggests the agent learned to recognize and approach the sticker before achieving reliable removal
• Agent achieves approximately 70% sticker-removal success rate by end of 500k training stepsfinding0.769
  Main behavioral result demonstrating the model's efficacy in the mirror-mark task
• The sticker-removal behavior induced by the self-prior corresponds to stimulus-elicited intention rather than endogenous intention, aligning with the developmental view of early intentional agencyclaim0.766
  Connects the model's behavior to Zaadnoordijk and Bayne's taxonomy of intentional agency
• Passing a sticker-bearing latent through the self-prior removes the sticker in reconstruction, confirming distribution favors sticker-free statefinding0.726
  Shows the self-prior's generative distribution rejects sticker-bearing states