INDEX

Explanations

reasoning and justification

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

ska

2.67

sal

2.38

czą

2.33

mals

2.23

ssymb

2.21

dering

2.18

shiro

2.17

ske

2.14

annya

2.13

harap

2.11

POSITIVE LOGITS

 grenade

2.68

ोनेशिया

2.51

 собственно

2.51

пература

2.50

 relinquish

2.47

 фай

2.46

разу

2.45

KindOfClass

2.44

 missile

2.41

 ไฟล์

2.40

Activations Density 0.005%