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 cater waiters bustled around, but Sarah barely noticed. The champagne, normally a pleasure, tasted like ash in her mouth. She could feel her father watching her, his smile bright, but she couldn't meet his gaze for more than a moment.

He had always supported her dreams, sacrificed for her, and she had repaid him with… what? She tried to grasp it and failed. She had let his calls go unanswered, her visits dwindle. Now, the weight of the realization crushed her.

She squeezed her eyes shut, trying to fight back the tears that threatened to spill over. Her heart ached with a dull, persistent throb. The joy of the evening was completely lost to her.

Emotion: regretful

Cluster: Shame / Guilt
PC1 (Valence): -1.62 Negative
PC2 (Disposition): 0.76

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
L0.801
S0.774
H0.513
悔0.479
been0.473
Suppressed:
de-0.407
--0.320
随后-0.317
(!)-0.310
oldukça-0.301