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 constant demands of the students, the administration, and her own self-imposed standards had begun to take their toll. Ms. Johnson felt as if she was walking through molasses, her every movement heavy, her every thought clouded. Sleep was a distant memory.

A small, brightly colored envelope had been slipped under her door. Inside was a hand-drawn card, depicting a cartoon version of herself, surrounded by happy students. A brief message was scrawled at the bottom: "You make a difference." It was signed by several of her former students, including one she had struggled to reach, Liam.

She looked at the card, the bright colors offering a surprising jolt of energy. A smile, small but genuine, cracked across her face. She felt a lightness, as though a heavy weight had been lifted from her.

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