INDEX

Explanations

permitted provided

This neuron activates on numeric literal tokens (particularly hexadecimal and other number constants).

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

 whose

-1.02

 from

-1.02

 even

-1.02

on

-1.01

 With

-0.94

 information

-0.91

 their

-0.90

 During

-0.88

祈祷

-0.88

sy

-0.87

POSITIVE LOGITS

 dasselbe

1.02

攙

0.94

 meistens

0.93

Cuadro

0.90

 barnet

0.90

oczes

0.90

żo

0.90

 pretrained

0.89

Neymar

0.88

や

0.88

Activations Density 0.001%