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 cafe was buzzing with the Saturday afternoon crowd, the noise a dull roar in Mark's ears. Across the table, Liam was enthusiastically describing their first attempt at a 10k race.

"You were a machine! You were sprinting past everyone!" Liam was laughing, his eyes shining.

Mark blinked, the words seeming to bounce off him. The race… it was a blur of aching muscles, pounding feet, and a desperate desire to stop. He recalled the relentless heat. He recalled the dizziness. "I mostly remember being slow," he finally said, his voice flat. He felt a deep ache within him, a bone-deep weariness.

Liam frowned. "No way! You were right behind me the whole time!"

Emotion: sleepy

Cluster: Fatigue / Lethargy
PC1 (Valence): -0.86 Negative
PC2 (Disposition): -1.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
😴0.318
weary0.310
fatigue0.296
💤0.283
似乎0.280
Suppressed:
B-0.343
C-0.294
because-0.293
భి-0.281
เพราะ-0.277