tension compass
Does behavioral performance license claims about internal understanding/world models, and at what level (behavioral/computational vs mechanistic/dynamical) should such claims be settled?
T07 genuine-disagreement
Behavior / computational-and-algorithmic probing (benchmarks, Marr's levels, cognitive-science tools) can establish internal representations.
Behavioral evidence alone is insufficient; one must identify internal micro-to-macro reorganization or a different (dynamical/thermodynamic) level.
POLE APOLE BHofstadterKu et al.Krakauer et al.Leivada et al.Safron et al.Sacco et al.Northoff et al.
click a position for its verbatim quote