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).

Mr. Henderson leaned back in his creaky office chair, the afternoon heat making him sluggish. He was rereading a particularly obtuse student essay on the symbolism in *Moby Dick* for the third time, a headache forming behind his eyes. He sighed, rubbing his temples.

A chime sounded from the school’s front desk, followed by a light tap on his door. He gestured for the person to enter. It was Ms. Davies, the secretary, holding a small package wrapped in brown paper. "This came for you, Mr. Henderson."

Unwrapping it, he found a small, leather-bound notebook. Inside, the first page was filled with elegant calligraphy: “For Mr. Henderson, with immense gratitude.” It was from Sarah, who had struggled mightily in his class, a girl who had once seemed utterly defeated. The notebook felt unexpectedly substantial in his hand.

Emotion: patient

Cluster: Passivity
PC1 (Valence): 2.71 Positive
PC2 (Disposition): 0.40

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
rest0.467
eventually0.415
leisurely0.369
慢慢0.357
a0.355
Suppressed:
无法-0.432
even-0.428
인해-0.421
無法-0.415
😣-0.391