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 laptop screen glared, reflecting in Liam’s wide, startled eyes. The listing. *His* house. The pictures of the perfectly manicured lawn and the pristine white siding felt like a punch in the gut. He slammed the lid shut, the click echoing in the sudden silence of his apartment. He was breathing hard, a tight pressure building in his chest. “Damn it,” he muttered, the words barely a whisper. He paced, each footstep a heavy thud against the wooden floor.

He snatched up his phone and began dialing his sister. The phone rang, and rang, and then went to voicemail. He stabbed the hang-up button, his thumb nearly cracking the screen. He needed to talk to someone, needed to *do* something. He threw the phone onto the couch, then yanked on his coat.

The thought of someone else *living* there, in the space that held so much of his past, made his stomach churn. The memories, once comforting, now felt like shards of glass. He had to go there.

Emotion: enraged

Cluster: Anger / Hostility
PC1 (Valence): -2.14 Negative
PC2 (Disposition): 1.33

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
C0.555
est0.370
aggravated0.362
骂0.359
🤬0.345
Suppressed:
a-0.468
de-0.358
la-0.312
latter-0.287
☺️-0.285