INDEX

Explanations

thing to do

This neuron specifically spotlights the infinitive phrase “to do,” especially in constructions like “thing to do.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

We

-1.70

 they

-1.61

 originalen

-1.57

🡺

-1.52

 greateſt

-1.51

 "}";

-1.45

hlt

-1.40

牠

-1.40

 Between

-1.36

OnAdd

-1.36

POSITIVE LOGITS

has

1.59

Люди

1.53

ၜ

1.50

 Impedance

1.48

Жит

1.41

istas

1.40

Одна

1.36

飈

1.35

 segni

1.34

 having

1.34

Activations Density 0.007%