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).

From the moment she read the notice, Clara found it difficult to focus. Her morning coffee tasted like ash. The buzzing of the refrigerator felt like a personal affront. She found herself staring blankly at the computer screen, the words swimming before her eyes.

She tried to do a simple task, like ordering groceries online, but each click felt heavy, each form a monumental task. The thought of all the hidden tasks, the endless details she would need to attend to, felt like a weight crushing her chest. Her muscles felt tight, and she had a constant headache.

She abruptly closed her laptop and walked to the window, the city lights reflecting in her wide, frightened eyes. The world looked immense and menacing.

Emotion: overwhelmed

Cluster: Fear / Anxiety
PC1 (Valence): -2.52 Negative
PC2 (Disposition): -0.94

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.385
😫0.332
😖0.319
😣0.319
😩0.314
Suppressed:
la-0.576
de-0.496
a-0.297
happy-0.290
l-0.264