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 grand cathedral towered above him, its stained-glass windows depicting scenes of fire and brimstone. He and Maya had been children then, accompanying his parents on a trip to Europe. He’d never forgotten the visit.

“It was so intimidating, wasn't it?” Maya said, her voice soft. “I was so scared, I hid behind you the whole time.” He blinked. His memories were of a sense of wonder, of being spellbound by the architecture, of walking through shadows.

He didn't recall her fear, he didn't recall hiding. He had thought that they were in thrall to the beauty of the place, together. He stood frozen, the magnitude of the building suddenly pressing down on him. His fingers traced the cold stone. A coldness spread through him that had nothing to do with the temperature.

"No, I…" He began, then trailed off, unsure of how to voice the sudden, bewildering gap. He was silent.

Emotion: awestruck

Cluster: Surprise / Confusion
PC1 (Valence): -0.17 Negative
PC2 (Disposition): -1.20

Role in Research

This story is one of 1,000 stories generated for the emotion awestruck. 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 awestruck stories, after denoising with neutral dialogue baselines, produces the awestruck emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

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

Promoted:
L0.271
suddenly0.260
stranger0.252
plötzlich0.246
wonder0.242
Suppressed:
la-0.450
싶-0.350
K-0.319
rest-0.248
싶은-0.238