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 office was thick with the scent of stale coffee and impending doom. He watched as the new hire, Thomas, bounced in, full of youthful energy. He felt an unwelcome squeeze in his gut.

"So, Thomas, let's get you acquainted with the website," he mumbled, his voice gravelly. He began a slow, deliberate tour of the online platform, skipping the most relevant sections.

He spoke about the complicated coding, dismissing the basic tasks. He purposely made the process seem intricate, confusing, and needlessly complex. Thomas kept nodding, clearly trying to absorb everything.

His hands were clenched in his lap. He felt a wave of nausea. He took a deep breath, trying to regain control. He would make it as difficult as possible.

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