INDEX

Explanations

applications of

The neuron detects mentions of “application” (especially “applications”) in the text.

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

горь

-1.13

GOTREF

-1.13

щению

-1.10

似た

-1.08

Ỗ

-1.01

 المسؤول

-1.00

 llegamos

-0.99

 radikal

-0.98

 llevaba

-0.98

ónica

-0.98

POSITIVE LOGITS

of

1.75

 uses

1.64

use

1.64

 applications

1.51

 purposes

1.47

Uses

1.42

for

1.34

용

1.30

 Uses

1.27

 usage

1.23

Activations Density 0.100%