INDEX

Explanations

go followed by a word

The neuron primarily detects the exact token “go.”

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

there

-2.33

इटम

-2.19

 noting

-2.06

when

-1.95

ある

-1.92

ARIOS

-1.91

<td>

-1.89

at

-1.87

琊

-1.82

by

-1.81

POSITIVE LOGITS

2.70

erapeutics

2.34

ències

2.27

ꦧ

2.25

渃

2.17

maline

2.17

 Coruña

2.00

 senseless

1.97



1.96

enital

1.92

Activations Density 0.017%