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 doorbell’s shrill ring sliced through Eleanor’s attempt at peace. She’d been enjoying a cup of chamomile tea, her only pleasure after a day of spreadsheet purgatory. Now, here she stood, fists clenched, staring at a bouquet of lilies. Lush, fragrant lilies. Without a damn card. "Wonderful," she muttered, snatching them up and slamming the door.

She tossed the offending flowers onto the kitchen counter. The scent, previously a subtle elegance, was now a nauseating assault. She ripped open the fridge, grabbed a bottle of sparkling water, and chugged it down, the bubbles a minor distraction. Who sent these, and why couldn’t they be bothered to say? Her jaw was tight, and she felt a tremor in her hands.

The vase hunt began, a frantic search that ended in a chipped ceramic monstrosity. As she forced the stems into place, she felt a flash of heat behind her eyes. This was the third unexplained floral delivery this month. She would find out who was behind this needless act of… something.

Emotion: irritated

Cluster: Anger / Hostility
PC1 (Valence): -0.86 Negative
PC2 (Disposition): 1.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.370
S0.282
P0.264
😠0.248
aggravated0.233
Suppressed:
own-0.300
joyful-0.211
joyous-0.207
खोले-0.206
newfound-0.205