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 company picnic was supposed to be fun. However, the cheerful atmosphere grated on me. As I listened to the small talk, I discovered, in a casual conversation, that the new hire was now making more. I immediately lost my appetite, pushing the burger around my plate.

I excused myself early, retreating to a quiet corner, the joyous sounds of the party fading into the background. The sunshine felt oppressive, the laughter hollow.

I started to focus on my flaws. I felt inadequate and unworthy. The feeling was a heavy weight on my shoulders. I felt my face flush with a mix of shame and anger.

I avoided eye contact with the new hire for the rest of the picnic. I did not speak to anyone. The picnic, which was supposed to be a good time, had become a source of profound discomfort.

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