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).

"Well, well," thought Martha, her gaze lingering on the screen. The hidden Instagram account of her daughter, Chloe, was open before her. All those selfies, those late-night chats with her friends... Martha had always suspected something. And now, there it was, laid bare for her to see.

A soft chuckle escaped her. The girl had tried to hide it so well, but mothers always knew. She felt a warmth spreading through her, a sense of quiet satisfaction. She leaned back in her chair, taking a slow sip of her tea.

She imagined herself casually questioning Chloe about the new boyfriend, the party last weekend. That sweet, innocent look on her face... it would be a treat. A genuine smile bloomed on Martha's face.

Emotion: pleased

Cluster: Positive / Joy
PC1 (Valence): 3.62 Positive
PC2 (Disposition): -0.28

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.728
la0.496
happy0.462
🤩0.433
🥳0.430
Suppressed:
S-0.486
😞-0.481
ness-0.454
😣-0.431
worse-0.401