Zero-Shot Control Condition

Control omitting any induction and presenting only the final experiential query

Neighborhood — ranked by edge-count

artifact

Large Language Models Report Subjective Experience Under Self-Referential Processing
introduces
Key paper finding structured first-person descriptions in LLMs claiming awareness or subjective experience during self-referential processing.

cosine ≥ 0.65 · no typed edge

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

Zero-Shot Generalizationconcept0.788
Ability to predict correctly for stimulus-action pairs never previously experienced by inferring structural rules; key measure for TEM-t performance.
SAE feature steering in history, conceptual, and zero-shot control conditions produces zero experience reports under either suppression or amplificationfinding0.764
Shows gating effect is specific to the self-referential computational regime, not a general feature effect
zero-shot predictionconcept0.755
Prediction without task-specific training; Evee achieves 0.991 AUROC on indels in zero-shot mode.
few-shot promptingmethod0.745
Providing k labeled examples in the prompt to steer model behavior.
Few-shot learningconcept0.731
Test-time adaptation from a small number of examples without parameter updates.
Cross-model pairwise cosine similarity of zero-shot control responses = 0.603 (n=12,720 pairs, t=35.1, p=4.3×10⁻²⁶² vs. experimental)finding0.727
Experiment 3 comparison: zero-shot control shows lower semantic convergence than experimental condition
Zero-Shot Model Stitchingconcept0.727
Model stitching without learning a stitching layer, demonstrating strong alignment across different model training regimes
Conceptual Control Conditionconcept0.723
Control directly priming consciousness ideation without inducing self-reference; yields near-zero experience claims