INDEX

Explanations

cooler, Pelican, Crucible

The neuron is triggered by occurrences of the word “cooler.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

at

-2.47

 Skilled

-2.45

on

-2.44

 unspeak

-2.27

 There

-2.22

and

-2.22

 indescri

-2.20

 incred

-2.19

You

-2.14

 wirtschaft

-2.08

POSITIVE LOGITS

);

2.83

 also

2.17

CurtirCurtir

2.11

珖

2.11

掉的

2.11

2.09

2.02

 pudieron

2.02

iding

2.00

༈

1.98

Activations Density 0.001%