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’d always envisioned himself graduating from Crestwood University, a name that echoed with prestige. Now, the glossy brochure in David's hand felt like a lie. The news had dropped like a bomb during his morning lecture.

He wandered the campus, the usual bustle now muffled, almost spectral. The laughter of students, once a cheerful soundtrack, grated on his ears. He kicked a loose pebble across the quad, a small act of defiance against the inevitable.

He sat on the steps of the administration building, his back curved like a question mark. The sun, ordinarily a source of warmth, felt cold against his skin. Each brick, each window, each tree on campus whispered a lament.

Emotion: sad

Cluster: Sadness / Despair
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.436
πŸ˜”0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322