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).

Rain lashed against the windows, mirroring the tempest within Ethan. He hadn’t touched the garden tools since… well, since everything changed. The once vibrant plot was a tangle of weeds and forgotten dreams. A polite tap on the door broke through his despondency. Mrs. Gable, a neighbor with a perpetually sunny disposition, stood shivering on his porch.

"Ethan, dear," she said, her voice unusually subdued, "I was hoping, perhaps, to cultivate a small herb garden. Could I, by any chance, use a strip of your yard along the fence?" He looked at the vast expanse of grass, the neglected soil.

He found himself shaking his head, a gesture he could not explain. The idea of someone else's hands reaching for life when his own felt so useless felt like another painful reminder. He stammered a refusal, his face burning, wishing she’d just go away and leave him in peace.

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