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 train ride felt endless. Ava stared out the window at the blurred landscape, the rhythmic clatter a dull, soporific lullaby. The job application, the emails, the endless revisions – it all felt like a marathon she’d already run. Her friend, Chloe, was going for the same position, and Ava knew Chloe’s relentless energy was going to give her a run for her money.

She felt a strange ache in her bones as she finally disembarked, the sudden movement causing her head to swim. The city bustle felt overwhelming; the thought of engaging in a conversation felt impossible. She checked her phone and saw a text from Chloe. "Good luck!" it read. Ava swallowed hard, and started the long walk to the interview.

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