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 gallery felt vast and echoing, the vibrant artwork mocking her muted despair. Sheโ€™d always envisioned this day, her first solo exhibition, with her mentor, Liam, at her side, beaming with pride. Instead, she stood alone, the opening night a silent, solitary vigil.

She ran her hand over a canvas, the texture of the paint a cold comfort against her skin. The last message she'd received from Liam was an automated out-of-office response, a generic acknowledgment of his departure. No personal message, no congratulatory note. The knot in her stomach tightened.

She caught her reflection in the glass of a painting. Her face looked pale, her eyes vacant. She was surrounded by her work, a testament to her talent, but the achievement felt hollow, devoid of meaning. She felt a profound disconnect, a deep sense of being utterly alone in a crowded room.

Emotion: worthless

Cluster: Sadness / Despair
PC1 (Valence): -2.10 Negative
PC2 (Disposition): -0.15

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.427
S0.405
๐Ÿ˜ž0.386
ness0.339
็ผบไน0.326
Suppressed:
de-0.320
(!)-0.286
้šๅŽ-0.268
๐Ÿ˜-0.262
tc-0.255