INDEX

Explanations

The neuron consistently lights up on the word “collapse” (including in compound tokens like the CSS property “border‐collapse”).

collapse

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

-2.64

蜮

-2.31

躥

-2.30

 lentamente

-2.27



-2.22

 フォント

-2.17

The

-2.14

 exemplified

-2.11

῞

-2.08

豨

-2.06

POSITIVE LOGITS

↵

3.42

with

2.58

){

2.47

</sup>

2.17

2.05

!!!!

2.00

 avec

2.00

歃

1.98

 Eigentümer

1.95

</h4>

1.94

Activations Density 0.010%