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 ornate living room, usually a source of comfort, now felt oppressive. Across the room, Oliver, her brother, was practically vibrating with excitement. He'd always been the one who chased the light. She, on the other hand, felt a dull ache that seemed to seep into her bones. The solicitor’s voice, a monotone drone, was barely audible above the hum of the air conditioning.

"To Oliver, the art collection." Oliver’s grin was blinding. The collection, a treasure trove of paintings and sculptures, was a fitting inheritance for his refined tastes.

"And to Clara," the solicitor went on, "the photograph album." Clara stared at the antique book with a binding of faded leather. Its pages held old photographs, a record of the family's past. She felt something twist inside her chest. The album seemed to mock her, full of people she barely knew, a legacy of a life she couldn't understand.

Emotion: gloomy

Cluster: Sadness / Despair
PC1 (Valence): -1.68 Negative
PC2 (Disposition): -1.37

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.368
无法0.319
😔0.300
いつも0.270
쓸0.255
Suppressed:
la-0.432
l-0.419
de-0.398
B-0.294
!-0.259