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 email from *The Collegiate Review* sat in Amelia’s inbox, unread, for three days. She knew what it contained: the acceptance of her submission, the essay she'd poured her heart into about her grandfather’s woodworking. She’d always wanted to be published. But the thought of it, of being *exposed*, triggered a slow burn of unease that coiled in her stomach. She sat at her desk, the open window letting in the humid afternoon air, and finally clicked the message. The review editor’s words were kind, but the accompanying PDF was the kicker. There, on page six, nestled between essays about climate change and social justice, was *hers*. A deep, slow breath escaped her, and she smiled slightly, picking up her coffee cup and taking a long sip. The moment was hers to savor.

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