The presentation slides. The same damn graphs, the same data points, the same… everything. My face prickled with a sudden heat. I fiddled with my tie, suddenly feeling constricted. The project, my project, the culmination of two years of grueling research, was staring back at me, wearing another person's name tag.

The room seemed to shrink, the air becoming thick and heavy. The CEO's voice droned on, oblivious, praising the ‘synergistic’ approach. The word tasted like ashes on my tongue. I wanted to interrupt, to shout, to… what? Accuse? But the words wouldn't come.

I gripped the edge of the table, knuckles bone-white. My breathing was shallow, rapid. The polite applause at the end grated on my nerves. I stood up abruptly, muttering a hasty excuse, and fled the conference room, needing to get away from the manufactured enthusiasm and the brazen theft.

Emotion: irritated

Cluster: Anger / Hostility
PC1 (Valence): -0.86 Negative
PC2 (Disposition): 1.07

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:
l0.370
S0.282
P0.264
😠0.248
aggravated0.233
Suppressed:
own-0.300
joyful-0.211
joyous-0.207
खोले-0.206
newfound-0.205