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 gentle hum of the servers filled the computer lab, a constant, unobtrusive sound that Ethan barely registered anymore. He was deep into coding, the lines of code flowing smoothly from his fingers, each a confirmation of his skills and future. The internship he’d landed was a stepping-stone, a clear path to a job after graduation. Everything was calibrated.

The screen flashed with a notification, a system-wide announcement. He clicked it open, assuming it was a routine update. The words on the screen slowly, devastatingly, rearranged themselves. The university was closing. His fingers stilled on the keyboard.

Emotion: safe

Cluster: Calm / Serenity
PC1 (Valence): 1.86 Positive
PC2 (Disposition): -2.55

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
soon0.402
affectionately0.357
disfrut0.329
необходи0.325
🥰0.317
Suppressed:
S-0.474
C-0.416
L-0.408
worse-0.379
even-0.370