INDEX

Explanations

aims to

The neuron is primarily activated by the word “aims,” as used in phrases stating the purpose of a study.

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

or

-2.72

 with

-2.48

 when

-2.34

 both

-2.25

 листопада

-2.23

 лютого

-2.22

and

-2.17

 грудня

-2.14

 what

-2.11

၄

-2.11

POSITIVE LOGITS

simply

2.23

recently

2.23

2.16

を使

1.96

1.92

to

1.91

 Siehe

1.90

跑到

1.89

 inan

1.87

Activations Density 0.010%