nucleation

THEORY-1224: In Claude language models, projecting out the most active J-space directions in early workspace layers lowers the rate of felt, first-person, sensory language in the models' narrations of their own processing and in their descriptions of another person's experience, while norm-matched perturbations along other directions do not and the text stays coherent; so in these models the experiential register of report depends on the access-like workspace and is not specific to self-report

1 min read

Properties4
StatusProposed
Tagsconsciousness, cognition, representation-learning
SourceLIT-1347 — Verbalizable Representations Form a Global Workspace in Language Models
ExtendsTHEORY-869 — In Claude Sonnet 4.5, Opus 4.5 and Opus 4.6, Gurnee, Sofroniew et al. found that projecting out the ten most active J-space directions in layers 38 to 54 of 100 lowered an experiential-language score, graded by Opus 4.6-based model graders on five or six prompts per condition, both in the models' narrations of their own processing and in their descriptions of another person's experience, while norm-matched perturbations along other directions left it near baseline and the text stayed fluent; on Haiku 4.5 coherence failed first

Graph View

Source

LIT-1347, read in full in NOTE-1069. What the source measured is THEORY-869, which this entry extends.

The claim

The authors conclude from the ablation that the experiential register of these models’ reports depends on the J-space, their access-like workspace, and that the dependence is the same for reports about others, so it is not specific to self-report.

What would settle it

The promote_when, carried word for word from THEORY-869 when that entry was narrowed. More readings of the same source cannot settle it.

Backlinks