What question did this study set out to answer?

The aim is to explore how well a cognitive architecture can accurately describe its own internal states.

July 28, 2026Open Access

Grounded Self-Report in an Integrated Cognitive Architecture: Accurate Description of Internal States as a Measurable Precursor to Machine Self-Modelling

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial examines internal state accuracy in machine self-reporting, suggesting improved cognitive modeling.

The aim is to explore how well a cognitive architecture can accurately describe its own internal states.
Developed an independent cognitive architecture called XK82.
Trained the system to produce language and describe its internal states without prior weights.
Conducted trials to assess accuracy in producing self-descriptions and contrasted them with baselines.
The system accurately described internal states in 91.7% of trials (22/24), far exceeding baseline performance of 41.7%.
An independently scored internal variable also reached 66.7%.
The decision boundaries were placed within 0.01 of empirical thresholds, indicating significant precision.

Junai P. Felix (2026) studied this question.

View Full Paper

Ask AI

Bookmark