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 box smelled of lavender and paper. Amelia’s fingers trembled as she pulled out the first assignment: a book report on *Little Women*. Her mother had been a meticulous student, a fact that had always seemed distant and unimportant. Now, as Amelia read her mother’s youthful handwriting, a soft joy bloomed within her.

She had always believed their bond was fragile, easily broken. This, however, felt strong, a tangible connection forged across time. The world outside the box seemed to quiet.

She carefully stacked the papers, each one a precious fragment of a life she was just beginning to understand. A sense of anticipation swelled, filling her entirely.

Emotion: hope

Cluster: Positive / Joy
PC1 (Valence): 3.59 Positive
PC2 (Disposition): -0.12

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.065
la0.865
B0.465
🤩0.412
H0.407
Suppressed:
😞-0.459
C-0.451
不-0.449
😣-0.443
S-0.442