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 air in the elevator was thick with his own nervous energy. He ran his fingers through his hair, smoothing down the spikes. The name on the directory, "Olivia Pierce, PhD," glared at him. He hadn't seen her in twenty years, but the memories of her sneering face were as fresh as yesterday.

He swallowed hard as the elevator doors opened. Each step down the hallway amplified the sound of his own heartbeat. The building was quiet, sterile, each footstep an echo of the past.

He reached the door, hesitating. He could turn back, make an excuse. But curiosity, or perhaps something darker, kept him rooted in place. He knocked, the sound like a drumbeat of doom.

Emotion: suspicious

Cluster: Suspicion / Vigilance
PC1 (Valence): -0.82 Negative
PC2 (Disposition): 0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
🤔0.420
🤔0.415
🧐0.362
疑0.352
suspicious0.352
Suppressed:
and-0.307
alleviate-0.301
ได้รับการ-0.298
allevi-0.285
mouthful-0.284