INDEX

Explanations

html tags or ui components

The neuron detects numeric tokens (i.e. any token containing digits, such as years, version numbers, measurements, or citation numerals).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ंबर

0.60

 herringbone

0.59

畽

0.58

 sien

0.57

 côtes

0.55

 decía

0.54

 कर्त

0.53

 aisles

0.53

 velké

0.52

芯

0.52

POSITIVE LOGITS

<>

1.15

<>

1.04

<div>

0.85

<html>

0.73

div

0.73

div

0.68

try

0.63

<body>

0.63

<>(

0.63

 Container

0.59

Activations Density 0.001%