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 harsh fluorescent lights of the shop seemed to amplify the silence. Daniel, leaning against a display case, hadn't slept well in weeks. The news from the hospital had hung over him like a suffocating shroud. He felt leaden, his limbs heavy with a weariness that went beyond physical exhaustion.

He was examining a tarnished silver locket when something at the back of a case caught his eye. It was a ring, a simple band with a small, unassuming pearl. It felt achingly familiar. He asked the shop owner to unlock the case. He turned the ring over in his calloused hands, reading the engraving - "Forever yours, Arthur".

He closed his eyes for a moment, the weight of the realization pressing down on him. A single tear escaped and trailed down his cheek.

Emotion: melancholy

Cluster: Sadness / Despair
PC1 (Valence): -1.78 Negative
PC2 (Disposition): -1.58

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.445
😞0.376
😔0.337
いつも0.327
sadness0.312
Suppressed:
de-0.739
la-0.548
l-0.497
/-0.314
!-0.290