INDEX

Explanations

CSS properties and titles

The neuron is selectively activating on numeric tokens (numeric literals and measurements) 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

 uska

-1.20

และ

-1.17



-1.13

 tien

-1.11

 ayur

-1.09

 unſ

-1.06

甶

-1.04

 FRANK

-1.02

 dezelve

-1.02

 PERSONAL

-1.02

POSITIVE LOGITS

in

1.41

as

1.40

for

1.38

 Without

1.34

 With

1.31

 from

1.29

on

1.27

 when

1.26

 Despite

1.25

at

1.23

Activations Density 0.004%