INDEX

Explanations

symbols and punctuation in math

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wiÄħ

-0.09

æ±¶

-0.08

illy

-0.08

apun

-0.08

uder

-0.08

è¾¾åħĭ

-0.07

 Estimated

-0.07

uhn

-0.07

ÃŃst

-0.07

ÏĢÏĮÎ½

-0.07

POSITIVE LOGITS

 vice

0.07

letter

0.07

god

0.06

plus

0.06

 correct

0.06

ç»į

0.06

åıįä¹ĭ

0.06

éģ¨

0.06

andi

0.06

vice

0.06

Activations Density 0.048%