mad - Story 87386 | Emotion Vectors in Gemma4-31B

The gym felt like a pressure cooker. Sweat dripped into Ben's eyes, blurring his vision. He pushed the weight, his muscles screaming, each rep a furious stab at the betrayal he felt. This wasn't supposed to happen. He had trusted her, thought he had found someone special. Now he was in a brutal contest with another man.

He slammed the weight back onto the rack, the clang echoing in the otherwise quiet gym. He felt a surge of adrenaline, his heart pounding against his ribs. The anger churned in his stomach. He wasn't going to show up to the cafe where they’d agreed to meet him and the other man.

He finished his workout and strode to the locker room, his movements sharp and deliberate. He needed to get away, to escape this suffocating feeling. He wouldn't pick up her calls. He wouldn't answer her texts. He would disappear.

Emotion: mad

Cluster: Anger / Hostility

PC1 (Valence): -1.94 Negative

PC2 (Disposition): 1.26

Role in Research

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

Logit Lens (Layer 40)

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

Promoted:

`C`	0.448
`骂`	0.344
`est`	0.340
`🤬`	0.336
`aggravated`	0.332

Suppressed:

`a`	-0.495
`de`	-0.485
`la`	-0.387
`H`	-0.344
`latter`	-0.290