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 air in the room felt thick, charged with unspoken truths. His breath hitched as he stared at the document, the numbers swirling before his eyes. He'd been looking for a phone charger and found a digital rabbit hole. The account, belonging to his wife, Maria, was filled with funds he never suspected.

He sank back into the chair, the sudden weight of the discovery pressing down on him. A slow smile spread across his face, followed by a series of quiet, almost involuntary chuckles. He thought of Maria's meticulous planning, her knack for finding bargains, her quiet strength.

He got up from the chair. He felt the need to get outside, to walk, to breathe. He threw on his jacket and headed for the door, a peculiar spring in his step.

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