INDEX

Explanations

The main thing this neuron does is detect references to an entity’s start or founding—especially words like “launch,” “inception,” “founding,” and adjacent years.

since its establishment

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

 nachdem

-1.20

-1.16

 will

-1.15

and

-1.14

んだり

-1.11

ۢ

-1.09

ご了承

-1.04

 предназна

-1.03

 protože

-1.02

 kemudian

-1.01

POSITIVE LOGITS

in

1.61

 last

1.24

kā

1.19

 linge

1.16

 years

1.12

 năm

1.10

lagu

1.10

 roce

1.08

дели

1.07

 alkoh

1.07

Activations Density 0.050%