INDEX

Explanations

license to use, copy, modify

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

 représentent

-0.92

 devront

-0.91

ReLU

-0.87

">"

-0.81

athan

-0.79

歺

-0.79

 compri

-0.79

roads

-0.78

UTA

-0.78

ATHAN

-0.78

POSITIVE LOGITS

its

0.82

Prez

0.81

 urged

0.77

 allows

0.76

 what

0.74

 المر

0.73



0.73

 examining

0.73

 moved

0.72

 brąz

0.72

Activations Density 0.039%