INDEX

Explanations

randomized controlled trial

This neuron identifies mentions of randomized controlled trials (e.g., “randomized,” “controlled,” “trial,” “group,” “intervention”).

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

-1.55

外界

-1.55

now

-1.50

大多

-1.50

 reposo

-1.38

 also

-1.36

नलोड

-1.33

 including

-1.32

 about

-1.30

 explain

-1.30

POSITIVE LOGITS

 ήταν

1.41

*}

1.41

 atlet

1.35

」。

1.31

 Privacidade

1.27

 solchen

1.26

 marché

1.26

 muzik

1.25

ちゃんが

1.24

"}";

1.23

Activations Density 0.236%