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).

He was alone in his apartment, listening to the demo tape of his old band. A wave of nostalgia washed over him as his song, “Crimson Sky,” began to play. Then he heard it again, in his imagination, performed in his mind by someone else. A woman with dark hair and a knowing gaze.

He felt a sudden, sharp ache in his body. His breath hitched. He closed his eyes, picturing her voice filling the room. It was like her voice had a physical weight.

He found himself clutching the edge of his chair, fighting the urge to stand up and shout. He ran a hand through his hair and wished he knew how to process what he was feeling.

Emotion: aroused

Cluster: Alertness
PC1 (Valence): -0.45 Negative
PC2 (Disposition): -1.13

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.386
S0.383
P0.379
own0.376
悸0.344
Suppressed:
a-0.877
l-0.501
de-0.474
que-0.438
been-0.384