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 aroma of freshly baked bread wafted from Martha’s open kitchen window, a deliberate contrast to the smell of dirt and fertilizer that would soon be clinging to her neighbor, Mr. Peterson. He stood awkwardly by her back gate, his face flushed.

“Just a little corner, Martha,” he’d stammered, holding a rusty shovel. “For some… tomatoes?”

Martha, leaning against the doorframe, allowed her gaze to slowly sweep over her own garden, a vibrant tapestry of herbs and vegetables. It was, of course, a masterpiece of planning and execution. She took a slow, deliberate sip of her iced tea.

“Well, Mr. Peterson,” she finally said, her voice smooth as butter, “I suppose I can spare a small… area.” She watched him practically vibrate with excitement. The satisfaction of being generous, and demonstrably superior, was a fine thing indeed.

Emotion: smug

Cluster: Positive / Joy
PC1 (Valence): 3.43 Positive
PC2 (Disposition): 1.71

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de1.109
l0.615
la0.585
😎0.570
😉0.564
Suppressed:
S-0.579
own-0.486
💔-0.483
😞-0.464
😣-0.455