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 dingy pub was the last place Amelia expected to hear her music. The band, a local group she'd never heard of, was tearing into her ballad, "Fallen Leaves." The singer, a woman with fiery red hair and a voice like molten gold, was pouring every ounce of emotion into the performance.

A tingling sensation raced up her spine. Her fingers tapped a rhythm on the table. She felt a powerful urge to clap, to cheer, to be a part of the moment. The weariness she'd been carrying around for months dissolved. She leaned forward, completely absorbed.

She bought the band a round of drinks afterward, and when she got home, she pulled out her guitar. The dust on it felt unfamiliar.

Emotion: rejuvenated

Cluster: Positive / Joy
PC1 (Valence): 3.32 Positive
PC2 (Disposition): -0.82

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
la0.547
de0.526
newfound0.523
invigor0.495
rejuvenated0.495
Suppressed:
😞-0.485
😣-0.451
不-0.434
worse-0.419
πŸ˜₯-0.385