INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

çĻ½é©¬

-0.26

chod

-0.25

iling

-0.25

è¹ļ

-0.24

è¶Ł

-0.24

-grow

-0.24

ToFront

-0.23

å·¡åĽŀ

-0.23

ettle

-0.23

 Schn

-0.23

POSITIVE LOGITS

quals

0.30

 first

0.25

atisf

0.25

ester

0.24

recall

0.24

çŃĭ

0.24

sumer

0.24

uC

0.24

iola

0.23

esting

0.23

Activations Density 0.889%

No Known Activations

This feature has no known activations.