INDEX

Explanations

world around

This neuron is primarily detecting the adverb “around” when it appears in text.

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

↵↵↵

-1.90

by

-1.72

 there

-1.64

 their

-1.59

↵↵↵↵

-1.44

↵↵↵↵↵

-1.43

 даже

-1.40

on

-1.38

</u>

-1.38

 الاطلاع

-1.37

POSITIVE LOGITS

1.61

which

1.43

1.37

 that

1.36

 politiet

1.34

that

1.34

ции

1.30

ляє

1.30

Activations Density 0.063%