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 concert hall buzzed with energy. Lights flashed. The crowd roared. His music, his scores, were being played by a prestigious orchestra, but he wasn’t the one who had given consent. Marcus sat in the shadows, hidden from view. He felt like a spectator.

The music swelled, the notes soaring through the air. He’d written it for himself, as a means of expressing the feelings he couldn't articulate. Now, it was someone else's triumph. Each crescendo, each diminuendo, felt like a deliberate jab.

He wanted to interrupt, to stop the music, to claim it as his own. But he remained frozen, consumed by a feeling of profound isolation. He felt the cold seep into his bones, and felt himself shrink away. He eventually slipped out, the music echoing in his ears.

Emotion: depressed

Cluster: Sadness / Despair
PC1 (Valence): -2.06 Negative
PC2 (Disposition): -1.08

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.417
每天0.353
无法0.330
own0.311
😔0.309
Suppressed:
de-0.492
B-0.393
P-0.352
l-0.310
!-0.307