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 screen flashed the dreaded word, and Sarah felt as though the world had tilted on its axis. "Delayed." The announcement, a cold, impersonal statement, shattered her carefully constructed composure. Her son's graduation ceremony, a culmination of years of hard work, was happening miles away, without her.

Her hands began to tremble. She gripped the edge of a seat, trying to steady herself. She'd promised him she'd be there. Sheโ€™d promised. The shame, and the frustration, burned like acid.

She leaned back, her eyes squeezed shut, picturing his face, his smile. She tasted a bitter wave of regret and disappointment. Each tick of the clock was another moment lost, another opportunity missed. The weight of it all was crushing.

Emotion: shaken

Cluster: Fear / Anxiety
PC1 (Valence): -2.72 Negative
PC2 (Disposition): -0.86

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.351
unable0.269
focus0.266
๐Ÿ˜ž0.258
๋‹ค์‹œ0.257
Suppressed:
de-0.769
la-0.579
a-0.375
happy-0.292
"-0.290