INDEX

Explanations

conflicting reports data opinions

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

⚌

-2.48

 eğer

-2.23

いただいた

-2.17

alitions

-2.09

 gilded

-2.02

to

-1.99

-1.95

 刺繍

-1.94

 fortitude

-1.94

 zweimal

-1.93

POSITIVE LOGITS

while

2.39

䛗

2.25

 そう

2.14

2.13

--"

2.11

but

2.06

2.05

/"

2.02

就像是

2.00

Activations Density 0.004%