INDEX

Explanations

code delimiters

The neuron detects website interface and navigational text, such as calls-to-action, labels (e.g. “Tags:”), prompts (“Next post:”), and phrases encouraging user engagement.

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

){

0.90

(){

0.82

(){

0.77

0.73

 throws

0.71

){

0.68

 thì

0.67

 führt

0.67

=>{

0.67

 //=>

0.67

POSITIVE LOGITS

butadiene

0.64

下記の

0.61

-}

0.61

-}\

0.61

ਨ੍ਹਾਂ

0.60

чайно

0.59

薄

0.58

below

0.57

等等

0.57

(*(

0.56

Activations Density 0.000%