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 familiar rhythm of the street, the sounds of children playing and cars passing, now felt distorted, like a broken record. She had always admired the beautiful flowers in Ms. Rodriguez’s garden, but now, knowing her secret fortune, the blooms felt like a betrayal. The knowledge was like a weight on her shoulders, a burden.

She avoided walking past the house, feeling a deep internal discomfort. The casual conversations now felt like a test, and she always failed. The smiles and the simple pleasures of their shared neighborhood felt like a joke at her expense.

Her own home, once a sanctuary, now felt cramped and inadequate. She began spending more time alone, the silence amplifying the quiet ache that had taken root in her chest. Everything felt unbalanced.

Emotion: guilty

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 1.23

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
S0.492
L0.398
탓0.354
been0.323
ashamed0.319
Suppressed:
de-0.400
(!)-0.279
la-0.276
😍-0.274
"!-0.264