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 found himself staring at the ceiling, unable to move. A knot had formed in his stomach, a tight, unrelenting pressure. The doctor's words echoed in his ears โ€“ โ€œThe records are mismatched.โ€ He tried to sit up, but his muscles seemed to have forgotten how.

He was supposed to be preparing for a presentation. The notes lay scattered on his desk, but his mind refused to focus. Each thought fractured, scattering into fragments. He ran a hand through his hair, tugging at the strands.

He tried to call his wife, his fingers clumsy on the phone. His voice, when he finally spoke, was a strained whisper. It was the same voice he used when he was in the dentist's chair. He didnโ€™t want to go to the dentist.

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