INDEX

Explanations

limiting factor

The neuron fires on the word “limiting” (as in rate-limiting, dose-limiting, limiting step, etc.), flagging when a constraint or bottleneck is being described.

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

térature

-1.53

ösung

-1.49

ativeness

-1.42

Literatur

-1.41

-1.40

 ребен

-1.33

ongi

-1.32

 ACKNOWLEDGMENTS

-1.30

cko

-1.30

⎆

-1.28

POSITIVE LOGITS

 with

1.65

 quiser

1.41

 further

1.35

 vostri

1.35

).

1.34

),

1.31

さんです

1.30

 privado

1.27

 limiting

1.26

unjungi

1.22

Activations Density 0.044%