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 relentless clatter had begun to seep into her dreams. She woke up tangled in the sheets, her heart racing. The house across the way was being reborn, but her own world felt permanently fractured. She felt a profound weariness.

She pushed herself to the window, watching the crew. Sunlight glinted off the tools, a sharp contrast to the gloom she felt. She thought of the bright colours that would eventually fill the newly renovated house, but those colours were a world away from her.

She shuffled to the kitchen and made some toast. She left it on the plate, untouched. The sound of a drill bit pierced the air, and she closed her eyes, wishing, once again, for silence.

Emotion: sad

Cluster: Sadness / Despair
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -1.50

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.436
😔0.405
S0.384
无法0.341
L0.337
Suppressed:
de-0.749
la-0.614
l-0.471
/-0.350
!-0.322