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 shadows in the living room offered a small, temporary refuge. Clara watched, from behind the curtains, her neighbor, Mr. Harrison, pacing back and forth. His impatient gestures spoke volumes. She drew the curtains completely, plunging the room into semi-darkness.

The thought of opening the door, of dealing with the inevitable conversation, filled her with a nameless dread. Her limbs felt heavy, leaden. Every muscle ached. The idea of the sun, of the outside world, felt like a threat.

She let the knocking continue, unanswered. She found herself staring at the patterns on the wallpaper, trying to find a distraction from the crushing weight of the day. The tree, his complaints, they all seemed irrelevant. The only thing she wanted was for everything to fade away.

Emotion: worn out

Cluster: Fatigue / Lethargy
PC1 (Valence): -1.12 Negative
PC2 (Disposition): -1.20

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😞0.265
😩0.234
η–²γ‚Œ0.226
que0.221
sighed0.221
Suppressed:
B-0.357
because-0.228
l-0.222
ΰ°­ΰ°Ώ-0.208
あくまで-0.208