INDEX

Explanations

prompts for action

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 retweet

1.19

 phosphorylation

1.14

 Reinh

1.05

zach

1.03

 hatırl

1.03

 println

1.01

erar

0.99

phosphorylation

0.98

tweet

0.98

ymax

0.98

POSITIVE LOGITS

 embarking

2.18

 attempting

2.17

 본격

2.06

试图

2.06

 приступи

2.01

开始了

1.96

 пыта

1.95

attempt

1.93

 venturing

1.92

 intentar

1.92

Activations Density 0.231%

No Known Activations