INDEX

Explanations

the start of something

The neuron fires on occurrences of the word “start” (primarily its past‐tense form “started”).

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

 وم

-2.63

-2.50

 нови

-2.50

 comprennent

-2.39

一

-2.31

 وأ

-2.30

 savvy

-2.28

 maravill

-2.27

thenburg

-2.27

POSITIVE LOGITS

闼

3.11

縂

2.83

诹

2.56

 ツリー

2.55

);

2.50

 breasted

2.41

,”

2.39

罽

2.38

 男士

2.34

Activations Density 0.023%