INDEX

Explanations

identifications or roles after 'a'/'an'

The neuron fires on the very first content token of a new block or heading (i.e. the first word of a paragraph, title, or section).

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

風味

-0.95

稜

-0.94

 একটি

-0.93

 then

-0.89

由于

-0.85

雫

-0.85

 personnels

-0.83

弔

-0.83

 början

-0.83

用来

-0.82

POSITIVE LOGITS

 member

1.60

方がいい

1.27

 part

1.27

 perigo

1.20

tad

1.19

為に

1.17

ただの

1.16

ほら

1.13

 victim

1.13

帰る

1.08

Activations Density 0.055%