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 smell of old paper and dust filled the tiny room. Liam clutched the faded, yellowed paper in his trembling hand. It was a treasure map, supposedly, he couldn’t shake the sense of being caught in some bizarre joke. He didn't know where it came from. The map promised a hidden treasure, a time capsule. He’d ignored it at first, discarding the paper, but now it was all he could think about.

The location: a forgotten corner of the sprawling university campus. He'd spent days studying the map, meticulously tracing the lines. Each step felt like a betrayal. The wind rattled the windows. As he followed the final clue, he felt a prickling sensation on his skin. He stopped, glancing back over his shoulder. He felt like he was being followed, that his every move was scrutinized, his thoughts exposed.

Emotion: paranoid

Cluster: Suspicion / Vigilance
PC1 (Valence): -2.37 Negative
PC2 (Disposition): 0.70

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
even0.418
마다0.378
paranoid0.371
afraid0.366
🕵0.360
Suppressed:
la-0.628
de-0.541
L-0.472
H-0.470
B-0.434