INDEX

Explanations

The neuron responds to tokens that are capitalized or in mixed‐case—e.g. proper names, month names, acronyms, and code identifiers/function names.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

strftime

0.62

much

0.57

 ощу

0.56

param

0.55

okedex

0.53

0.52

かもしれませんが

0.50

openqa

0.50

setParameter

0.50

слежи

0.49

POSITIVE LOGITS

()

1.98

(()

1.55

(()

1.53

()

1.48

():

1.43

 ()=>{

1.28

 ()=>

1.25

(){

1.23

(),

1.23

();

1.22

Activations Density 0.012%