Language models describe some internal states as good and others as bad. But whether models have a stake in them is an open question. Simply asking the model is unlikely to be informative. Any answer may be consistent with genuine introspection, superficial pattern-matching, or with fixed scripts learned in character training. We therefore study revealed preference. Rather than asking about a...
This paper advances a specific thesis about the relationship between consciousness and learning: namely, that the evaluative process central to learning—computing progress toward or away from goals—is identical to conscious experience. Valence, the positive or negative quality of experience, just is goal-relative prediction error. Viewed from the outside, this process is iterative optimization;...
Large language models sometimes produce structured, first-person descriptions that explicitly reference awareness or subjective experience. To better understand this behavior, we investigate one theoretically motivated condition under which such reports arise: self-referential processing, a computational motif emphasized across major theories of consciousness. Through a series of controlled...