INDEX

Explanations

code assignment operators

This neuron never activates—it doesn’t respond to any tokens and is effectively “dead.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

for

-2.33

at

-1.63

’

-1.63

 massively

-1.62

 where

-1.59

</h4>

-1.52

 Those

-1.51

on

-1.49

with

-1.43

 severely

-1.43

POSITIVE LOGITS

</i>

1.88



1.77

GEBUR

1.66

Begriffsklä

1.62

linawan

1.60

 heilig

1.60

馓

1.56

进入了

1.56

ᗪ

1.52

 indepen

1.52

Activations Density 0.002%