INDEX

Explanations

leading to

The neuron is specialized to detect occurrences of the verb “lead” (in forms like lead, leads, leading).

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

 untuk

-1.67

-1.54

-1.52

 using

-1.40

 aéreas

-1.39

-1.31

or

-1.30

openid

-1.30

ابي

-1.27

 aufnehmen

-1.27

POSITIVE LOGITS

 jusqu

1.69

athers

1.59

 Either

1.58

 Gibt

1.57

igth

1.51

It

1.49

他对

1.45

eer

1.45

ches

1.45

</sup>

1.45

Activations Density 0.009%