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 old man’s workshop was eerily quiet. Michael pushed open the heavy wooden door, the hinges protesting with a rusty squeal. He had expected to find Mr. Ito bent over a piece of wood, his brow furrowed in concentration, but the workshop was deserted. A small, handcrafted wooden bird, the last project they'd worked on together, sat perched on the workbench, a poignant symbol of what was lost. He picked it up, his fingers tracing its smooth curves.

His breath hitched. He had considered Mr. Ito a second grandfather. He felt a sharp pang, a hollowness in his chest. He turned slowly, searching the space as though he might somehow find his mentor hiding in the shadows. A profound sense of loss settled over him, heavy and suffocating.

Emotion: amazed

Cluster: Surprise / Confusion
PC1 (Valence): 0.74 Positive
PC2 (Disposition): -0.51

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤩0.301
de0.289
L0.284
😳0.251
அதிசய0.236
Suppressed:
ing-0.332
l-0.253
😞-0.240
unsuccessful-0.237
unsuccessfully-0.233