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 sterile white walls of the support group meetings were a prison. Mark sat hunched in a corner, avoiding eye contact. The platitudes, the gentle encouragements, the forced positivity – it was all a suffocating performance. He hated the pitying glances, the whispers of his name. He was not a victim. He was a survivor.

He drummed his fingers on the armrest, a restless rhythm against the saccharine sweetness of the speaker's voice. He was ready to leave when another person turned to him.

"You don't seem too… happy," a woman with tired eyes said.

He met her gaze, and found a flicker of recognition. "I'm not," he said flatly. "I'm here because I was told to be here."

"Me too," she replied, her lips twisting into a wry smile. "I'm Maya. Do you… have Chronos?"

He just stared at her, not answering, his shoulders stiffening in the closed space.

Emotion: defiant

Cluster: Defiance / Spite
PC1 (Valence): 1.11 Positive
PC2 (Disposition): 3.25

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.680
de0.668
l0.583
-0.524
either0.438
Suppressed:
own-0.423
似乎-0.404
熟悉的-0.401
不禁-0.394
难以-0.385