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 ring of the phone, late at night, in the room she wasn’t supposed to go into, had become a torment for Anna. The caller ID was always blocked. She finally saw the truth, on a utility bill, where the phone line was listed under his name. Her partner, Liam, was clearly hiding something.

She felt a wave of heat wash over her, her face burning. She walked to the window, the city lights a blurry, indistinct mass. She wanted to scream, to lash out, to break something. She gripped the cold windowsill, the pressure a small comfort.

She closed her eyes and took a deep breath. She knew, now, that the life she thought she had was gone.

Emotion: distressed

Cluster: Fear / Anxiety
PC1 (Valence): -2.39 Negative
PC2 (Disposition): -0.52

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😣0.304
😞0.296
worse0.291
😖0.285
😰0.276
Suppressed:
de-0.457
la-0.349
:)-0.239
B-0.232
happy-0.229