Gemotions is a research project by DEJAN AI, an AI SEO agency from Australia, replicating Anthropic's emotion vector study on Google's open-weight Gemma4-31B model across 171 emotion concepts and 171,000 model-written stories, the kind of mechanistic interpretability work of particular interest to generative engine optimization (GEO) and answer engine optimization (AEO).

The click of the car door echoed in the quiet street as she made her way, quickly, to the house. Grace hated walking at night. Every shadow seemed to hold a threat, every distant noise a potential danger. Sheโ€™d seen the letter sitting on the doormat earlier. She'd left it there. Sheโ€™d felt the chill wind blow past her. Now, she would have to face it.

The lights were on in the house, a beacon of sorts. She went through the motions, checking the security system, double-checking the locks. The envelope. She picked it up, her fingers brushing the paper, a cold dread creeping into her heart. She wasn't ready. Never would be.

Emotion: vigilant

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.15 Negative
PC2 (Disposition): 1.11

Role in Research

This story is one of 1,000 stories generated for the emotion vigilant. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 vigilant stories, after denoising with neutral dialogue baselines, produces the vigilant emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the vigilant vector is projected through the unembedding matrix.

Promoted:
๐Ÿ•ต0.354
๐Ÿ”0.320
๐Ÿ”Ž0.320
๐Ÿง0.316
scrutin0.310
Suppressed:
L-0.499
B-0.402
H-0.372
la-0.330
S-0.318