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 sunlight seemed to taunt her, streaming through the window as if to mock her inability to enjoy it. Sarah pulled the curtains shut, plunging her living room into a dim twilight. The hammering next door was relentless, a percussive torture.

She snapped at her husband when he offered a friendly greeting, her voice sharp and brittle. She retreated into her study, slamming the door. The sound still permeated the walls, a constant reminder of her helplessness.

Her hands trembled as she reached for a book, the words blurring before her eyes. The frustration was a knot in her stomach, a heavy weight that prevented her from settling. She needed quiet, she needed escape, and all she had was noise.

Emotion: upset

Cluster: Fear / Anxiety
PC1 (Valence): -2.30 Negative
PC2 (Disposition): -0.42

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.395
😞0.303
无法0.246
ness0.238
unable0.238
Suppressed:
de-0.533
la-0.457
l-0.223
😎-0.215
🤩-0.210