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 park, usually a place of joy and freedom, felt like a prison. The laughter of children, the squeak of swings, the barking of dogs, all grated on his nerves. He sat on a bench, head bowed, the weight of the rejection pressing down on him.

He’d spent countless hours here, visualizing plays, practicing his footwork, dreaming of the big game. Now, the field just seemed empty. The bright colors of the flowers felt garish. The sun, once a welcome comfort, now beat down relentlessly.

He pulled his baseball cap low, shielding his eyes from the world. He wanted to be invisible, to disappear into the anonymity of the crowd.

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぀も0.238
πŸ˜”0.230
:(0.228
Suppressed:
l-0.360
P-0.356
B-0.339
de-0.334
!-0.239