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 world felt muted. Everything was slow and blurry, like looking through a rain-streaked window. Liam hadn't been himself lately, always glued to his laptop, always distant. Sarahโ€™s attempts at connection were met with a distracted nod or a mumbled reply. She felt herself fading, a ghost in her own home.

One evening, while he was in the shower, she saw it: a password-protected file on his desktop titled "Project Echo." With trembling hands, she cracked the code - their anniversary date. She read the first sentence: "The woman I loved had become someone I no longer recognized."

She sat on the edge of the bed, a chill creeping into her bones. Her breath hitched, shallow and uneven. Her own name, repeatedly used within the text, stung with such profound rejection.

Emotion: dispirited

Cluster: Sadness / Despair
PC1 (Valence): -1.88 Negative
PC2 (Disposition): -0.98

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
๐Ÿ˜ž0.398
S0.348
โ€™0.325
ness0.320
ๆ— ๆณ•0.308
Suppressed:
de-0.297
l-0.256
๐Ÿ‘-0.251
sonra-0.240
aud-0.240