INDEX

Explanations

states of affairs

The neuron is looking for mid‐to‐long English content words (i.e. tokens of roughly six or more characters) that typically appear in dense, content‐rich prose.

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

ไม่มี

-1.32

voyez

-1.24

𓏸

-1.23

 construcciones

-1.23

 něj

-1.23

)—

-1.23

titudine

-1.23

masing

-1.17

","+

-1.17

âni

-1.16

POSITIVE LOGITS

of

1.63

for

1.16

in

1.14

voa

1.09

or

1.08

😤

1.07

了一些

1.07

1.06

 these

1.04

 lamang

0.98

Activations Density 0.093%