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 doctor's diagnosis had been a blow. Trauma from her childhood, it was a heavy weight, she was told. Seeing the drawings, the familiar shape of “Whisper” in her nephew, Tom’s, artwork, was, at first, a fresh wave of panic. But as she watched Tom, laughing and chatting with an invisible friend, something within her shifted. She felt her shoulders loosen, her breath deepen. She was not alone.

She sat next to Tom, asking him about Whisper’s adventures. A warmth bloomed in her chest, chasing away the cold fear. She reached out and traced the outline of the drawing with her finger, a small smile playing on her lips. She felt a lightness in her step as she walked out of the room.

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