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 review fluttered from Eleanor’s hand and landed on the cold marble floor. She didn't bother to pick it up. Her stomach twisted, a knot forming deep inside. "Mediocre... forgettable... lacking finesse." The words repeated like a cruel mantra.

She retreated to her small office, a space usually filled with the comforting aromas of herbs and spices. Today, the air felt stale. She sat at her desk, staring at the blank screen of her computer, the cursor blinking mockingly. She could barely breathe.

She thought about the critic. He was probably at some other restaurant by now, savoring a perfect dish, while she was here, deflated, trying to pick up the pieces of her shattered pride. She buried her face in her hands.

Emotion: scornful

Cluster: Anger / Hostility
PC1 (Valence): -0.76 Negative
PC2 (Disposition): 2.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.406
C0.391
S0.276
I0.268
愤0.243
Suppressed:
own-0.396
unexpectedly-0.288
Suddenly-0.260
soon-0.256
serendip-0.256