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 flickering candlelight cast elongated shadows across the dining room walls. Eva stared at the menu, written in what she assumed was Cyrillic. Dinner was at a restaurant, a new one that had opened near their apartment. Liam, her husband, had picked it.

He’d ordered in perfect Russian, something that felt like a revelation, or perhaps a betrayal. He knew she'd always struggled with her own family, their traditions, their culture. He knew she had never spoken Russian, nor did she plan to.

She felt a hollowness in her chest. A sensation that the air was too thin. She took a sip of her wine, its bitterness mirroring the twist in her gut. She no longer knew what he found in her.

Emotion: unhappy

Cluster: Sadness / Despair
PC1 (Valence): -1.62 Negative
PC2 (Disposition): -0.78

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.339
😞0.317
πŸ’”0.252
πŸ˜”0.249
😒0.241
Suppressed:
de-0.436
la-0.266
l-0.224
/-0.216
--0.199