INDEX

Explanations

programming language keywords

The neuron activates on tokens that end with an underscore (e.g. library or variable names with a trailing “_”).

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

for

-2.95

or

-2.67

 Taking

-2.36

at

-2.33

if

-2.28

 because

-2.25

as

-2.23

 after

-2.09

 spurred

-2.08

POSITIVE LOGITS

甞

2.34

 frança

2.34

鼍

2.31

uſ

2.30



2.25

鲧

2.14

 visse

2.13

 厂家

2.13

 kasa

2.11

 créer

2.09

Activations Density 0.026%