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 lyrics of the song spilling from the radio filled the room. Sarah tapped her foot, singing along as she painted her nails a vibrant shade of red. The afternoon sun bathed the room in warmth. She felt... light. The past few weeks with Tom had been like a dream.

Lately, though, Tom seemed… distant. He was glued to his phone, constantly texting, and taking calls in hushed tones. The explanation was always work. She'd trusted him, but suspicion, a unwelcome emotion, began to prick at her.

She had forgotten her phone in the living room. It buzzed and, without thinking, she picked it up. A message from “Babe” flickered across the screen. Her stomach plunged.

Emotion: happy

Cluster: Positive / Joy
PC1 (Valence): 3.98 Positive
PC2 (Disposition): -1.14

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.676
excitedly0.533
enthusi0.505
de0.498
🤩0.495
Suppressed:
S-0.896
😞-0.430
ness-0.410
それでも-0.387
L-0.379