INDEX

Explanations

abstract academic fields

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

thodoxy

-1.03

InThe

-1.02

 tärke

-1.01

 pizarra

-0.94

FILM

-0.93

もあった

-0.92

printer

-0.90

therners

-0.90

thodox

-0.87

booking

-0.87

POSITIVE LOGITS

 conditions

0.92

 very

0.91

 Logic

0.84

 idea

0.84

 like

0.82

CONDITIONS

0.82

 logic

0.81

so

0.80

Logic

0.80

 supposedly

0.80

Activations Density 0.054%