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).

Ms. Abernathy stared at the small, intricately carved wooden bird resting on her desk. It was a hummingbird, its wings seemingly frozen mid-flight. A note, penned in elegant cursive, lay beneath it: “Thank you, Ms. Abernathy, for teaching me to see the world. From, Leo.” A warmth bloomed in her chest. She picked up the bird, her thumb tracing the delicate carvings.

A sudden urge to reorganize her classroom took over. She’d been meaning to, for ages. Throwing open the window, she let in a blast of fresh air, the scent of spring blossoms. Chairs scraped as she rearranged them into a circle, ready for a discussion. Dust motes danced in the sunlight, and for the first time in weeks, the room felt vibrant, alive.

She hummed as she went, the hummingbird held in her hand, a tangible symbol of something she’d almost forgotten: the impact a teacher could have. She couldn’t wait to meet her new students.

Emotion: inspired

Cluster: Positive / Joy
PC1 (Valence): 2.98 Positive
PC2 (Disposition): 0.04

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
B0.563
invigor0.465
newfound0.446
هذ0.414
используя0.406
Suppressed:
😞-0.481
不-0.431
worse-0.428
😣-0.400
fact-0.389