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 coffee shop buzzed with a comforting energy, but for Daniel, the sounds blurred. Across the small table, Chloe was beaming. A new job, a new city, a chance to really spread her wings. He felt a sudden lightness in his chest.

“That’s…wonderful, Chloe,” he managed, reaching across the table to squeeze her hand. Her fingers were warm against his.

He started thinking about how they'd stay connected, the late-night video calls, the weekend visits. A renewed sense of commitment settled within him.

He pictured the two of them, years from now, sharing stories. He would learn all about her new city, he thought. He'd hear about her triumphs and failures. They would navigate life together, even from a distance.

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