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 heavy, as if gravity had intensified. He trudged through his day, each action a burden. His appetite had vanished. His sleep was fitful. Everything seemed to require far more effort than it should.

He wandered into a pawn shop, seeking… something. He wasn't sure what. Maybe a change of scenery. Maybe a glimmer of something to spark the embers of his interest. The cluttered interior offered no comfort. Then, he saw it. A flash of silver and diamond. The ring. His grandmother’s ring. The distinctive setting was unmistakable.

He asked the proprietor, a man with a walrus mustache, to see it. He turned the ring over in his hands, his fingers numb. The inscription, barely visible, sent a jolt through him: "Always & Forever." The air seemed to leave his lungs. He was caught in a vortex of memories.

Emotion: droopy

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.15 Negative
PC2 (Disposition): -1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.303
无法0.252
いつも0.238
😔0.230
:(0.228
Suppressed:
l-0.360
P-0.356
B-0.339
de-0.334
!-0.239