INDEX

Explanations

required actions and outcomes

The neuron activates on nouns or gerunds denoting formal actions, procedures, or directives (e.g. “instructions,” “precautions,” “restrictions,” “stealing,” “skills,” etc.).

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

も多く

-1.56

 cowok

-1.52

 lagerfeld

-1.44

Ꮣ

-1.38

為に

-1.38



-1.36

􏰀

-1.35

 ソー

-1.35

 alimentare

-1.34

 staffel

-1.30

POSITIVE LOGITS

as

2.05

on

1.89

in

1.78

to

1.75

 There

1.73

out

1.64

It

1.61

it

1.59

The

1.41

had

1.41

Activations Density 0.186%