INDEX

Explanations

service

np_acts-logits-general · gemini-2.5-flash-lite

references to vehicle or equipment maintenance and servicing.

oai_token-act-pair · claude-4-5-sonnet Triggered by @bfgcn7frf5

The main thing this neuron does is find mentions of “service” (and its variants like “serviced,” “servicing,” etc.), especially in maintenance or repair contexts.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

to

-2.16

-1.73

 Their

-1.56

 Neither

-1.41

 Those

-1.39

 Both

-1.38

 cuyas

-1.38

 that

-1.37

 Might

-1.36

-1.35

POSITIVE LOGITS

。”

1.64

 chimie

1.62

}$.

1.53

谔

1.53

𓐍

1.48

Jetzt

1.43

ximation

1.38

”，

1.38

Cevapla

1.36

 huset

1.36

Activations Density 0.062%

service

references to vehicle or equipment maintenance and servicing.

The main thing this neuron does is find mentions of “service” (and its variants like “serviced,” “servicing,” etc.), especially in maintenance or repair contexts.

No Comments

No Known Activations

service

references to vehicle or equipment maintenance and servicing.

The main thing this neuron does is find mentions of “service” (and its variants like “serviced,” “servicing,” etc.), especially in maintenance or repair contexts.

No Comments

No Known Activations