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 late afternoon sun cast long, skeletal shadows across the living room. Sarah sat curled on the sofa, the remote control a useless weight in her hand. Channel after channel flickered past, none of them holding any appeal. A dull ache throbbed behind her eyes. She hadn't felt like this in… well, a long time. The house felt too quiet, the silence pressing down on her like a physical burden. She sighed, a small, deflated sound. Thinking of the bills and the rising cost of groceries, she decided to check the online bank statements, and that's when she saw it - a separate account, listed with her husband’s name only. Her breath hitched.

The numbers blurred before her, each digit a tiny stab of bewilderment. Why hadn’t he mentioned this? A surge of something unpleasant, akin to a cold trickle down her spine, made her sit up a little straighter. The balance was substantial.

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