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 frantic scrubbing of the walls had stopped. Sarah was exhausted, both physically and emotionally. The drawings, the endless sketches of the thing she used to call "Shadow," were a constant reminder of a nightmare that had haunted her. Then, her daughter, Emily, brought home the new picture from preschool. Sarah stared at the drawing, a clumsy rendition of the familiar figure. A wave of exhaustion washed over her. She sank onto the couch, her muscles slack.

“Mommy, is he nice?” Emily asked, her voice small.

Sarah reached out and gently ruffled her daughter's hair. "Yes, honey. He's very nice." The stale air in the room suddenly felt fresh. The knot in her stomach began to unravel. She felt a sudden urge to laugh, a wild, almost hysterical release.

Emotion: relieved

Cluster: Positive / Joy
PC1 (Valence): 2.22 Positive
PC2 (Disposition): -0.97

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.638
finally0.606
finalmente0.597
relieved0.561
enfin0.546
Suppressed:
l-0.443
:(-0.439
😣-0.429
S-0.388
😖-0.385