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).

He dragged himself into the new IT department, the door feeling like a heavy oak barrier. The air hung thick with the smell of cables and the low drone of servers. David ran a hand through his hair, already feeling the ache in his shoulders from the previous night's lackluster sleep. The transfer was unexpected, a lateral move, a change of scenery.

He had a meeting with his new manager, Sarah, in fifteen minutes. He glanced at his reflection in the darkened screen of a computer. He looked… drained. The bags under his eyes were prominent; they echoed the hollowness he felt inside. He took a deep, shuddering breath.

“David? Right on time. Come in, have a seat.” Sarah’s voice, sharp and energetic, was jarring. He forced a smile, the muscles in his face protesting. She gestured toward a chair, and he sank into it gratefully. The weight of all the new information was already settling on him.

Emotion: tired

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.92 Negative
PC2 (Disposition): -1.00

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
weary0.254
fatigue0.249
😴0.238
sighed0.237
barely0.230
Suppressed:
C-0.268
because-0.225
B-0.219
🤟-0.217
భి-0.214