INDEX

Explanations

behavioral analysis

The neuron is tuned to detect mentions of “behavior” (including its variants like “behaviour” and “behaviors”).

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

'</

-1.51

alties

-1.38

ician

-1.30

ſtion

-1.25

例如

-1.25

Matic

-1.23

hetta

-1.23

 дешевле

-1.23

ztes

-1.21

'=

-1.20

POSITIVE LOGITS

on

1.88

 patterns

1.64

 their

1.54

 choose

1.51

 behavior

1.49

 through

1.49

 chooses

1.49

 data

1.41

 trends

1.32

 different

1.30

Activations Density 0.088%