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).

Michael slammed the notice onto the table. It landed with a satisfying thud, the ink smearing slightly from the force. He leaned back in his chair, a low chuckle rumbling in his chest. “Finally,” he muttered to himself. He’d meticulously planned this, every step.

He rose from his chair, a surge of energy coursing through him. He felt… invigorated. He stretched his arms, feeling the tension melt away. The apartment, with its peeling paint and leaky faucet, had been a prison. Now, he was breaking free.

He pictured the stunned look on Mr. Henderson's face when he would go to the apartment and find it empty. Oh, that would be a sight to behold. He took a deep breath, savoring the moment. He had won. This was just what he needed.

Emotion: triumphant

Cluster: Positive / Joy
PC1 (Valence): 3.54 Positive
PC2 (Disposition): 0.19

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
de0.734
la0.571
triumphant0.562
🎉0.531
ecstatic0.530
Suppressed:
😞-0.503
S-0.483
😣-0.482
😰-0.457
worse-0.444