INDEX

Explanations

let's break down

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Supporters

1.55

maxY

1.54

 maximise

1.53

Favourite

1.52

所有人

1.50

 Betfair

1.49

MinValue

1.48

(&:

1.45

 campaigners

1.44

setItem

1.43

POSITIVE LOGITS

 intrusions

1.15

RoutingModule

1.10

 fooling

1.08

 disent

1.05

 recognizing

1.03

 Redirect

1.00

 redirection

0.99

 حادث

0.98

 detour

0.98

hook

0.97

Activations Density 0.020%