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 air thrummed with the buzz of a circular saw. Emily tried, and failed, to drown out the noise with a loud podcast. The constant interruption to her workday was frustrating, but her annoyance quickly morphed into another feeling.

Across the lawn, the Johnsons were adding a second story to their home. Emily had met them briefly; they seemed excited, but also a little… uncertain.

She thought about their young children, who now would be subjected to the noise. She remembered the disruption of her own home renovations. The sleepless nights, the mountains of paperwork, the way everything always took longer than expected.

Emily glanced at her own freshly baked bread, realizing she had an extra loaf. She decided to walk over and offer it. Maybe it could help them a little.

Emotion: sympathetic

Cluster: Compassion / Love
PC1 (Valence): 1.13 Positive
PC2 (Disposition): 0.75

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
H0.521
L0.421
B0.401
la0.338
sympat0.305
Suppressed:
iostream-0.334
简直-0.306
再び-0.305
chẳng-0.296
не-0.296