INDEX

Explanations

tasks and duties

The neuron selectively lights up on the main action verbs that kick off individual job‐duty or task statements (e.g. “Builds,” “Tests,” “Installs,” “Performs,” etc.).

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

こういった

-1.71

さっき

-1.70

 totiž

-1.65

Irlande

-1.60

なんとなく

-1.59

 dokonce

-1.58

 nemlig

-1.58

なんだか

-1.55

 mismas

-1.55

たまたま

-1.54

POSITIVE LOGITS

of

2.72

 with

2.20

on

1.81

at

1.75

 related

1.69

and

1.63

all

1.61

 first

1.58

for

1.45

 section

1.45

Activations Density 0.141%