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 realtor’s glossy flyer landed on her doormat like a poisoned dart. *Invest in Your Future!* Emily scoffed. Her future was already planned, thank you very much, and it did not involve forking over a small fortune to own a shoebox in a building she'd loved as a rental.

She grabbed the flyer and walked to her neighbour's, Mrs. Rodriguez. They had been friends for nearly ten years. Together, they read the flyer and felt the same bitter resentment, the same shared sense of being pushed around.

"We will not let them," Emily announced, her voice low but firm. The realtor would not be getting in here. Her jaw set, she started to formulate a plan. Every act of resistance, big or small, now held a new significance.

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