INDEX

Explanations

any questions

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bes

-0.10

/Form

-0.10

 conspic

-0.10

Cue

-0.09

wald

-0.09

NOM

-0.09

 Philipp

-0.09

agine

-0.09

adas

-0.09

/tree

-0.09

POSITIVE LOGITS

 questions

0.32

 Questions

0.24

 question

0.22

questions

0.21

Questions

0.17

 unanswered

0.17

 Question

0.16

 ques

0.16

 queries

0.16

 questi

0.15

Activations Density 0.042%