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 attic air tasted of dust and forgotten things. Rain lashed against the windowpanes, mirroring the turmoil churning within Amelia. She clutched the yellowing paper, her mother's handwriting a familiar ghost against the stark white of the list. "See the Northern Lights," it read, bold and untarnished, unlike the woman who had written it, and unlike Amelia herself.

A tremor ran through her, leaving a chill in its wake. She sank onto an old trunk, the metal cold against her denim-clad thighs. The silence of the attic pressed in, broken only by the rhythmic drumming of the storm.

"Climb a mountain," another item declared. Her throat constricted. Her mother had always loved the outdoors, a vitality Amelia had inherited, then somehow misplaced along the way.

She closed her eyes, and a single tear escaped, tracing a lonely path down her cheek. The list felt like a challenge, a beckoning towards a life she now felt incapable of living. The mountain felt impossibly high.

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