INDEX

Explanations

github and Richard Cordray

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

for

-2.00

訁

-1.94

 With

-1.94

 Regardless

-1.93

とはいえ

-1.89

 encompassing

-1.88

 decently

-1.87

 garnered

-1.86

 Obviously

-1.85

糅

-1.85

POSITIVE LOGITS

镚

1.97

 kapak

1.97

iy

1.84

 kredi

1.82

 stuks

1.82

 diki

1.81

nü

1.80

honestly

1.80

龶

1.79

seers

1.77

Activations Density 0.016%