INDEX

Explanations

let followed by a word

The neuron detects uses of the word “let” (often at the start of an imperative or exhortation) in commands or invitations.

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

坂

-0.84

exempt

-0.81

asily

-0.81

১

-0.81

you

-0.79

یم

-0.79

dodge

-0.78

gląd

-0.78

giz

-0.78

роде

-0.75

POSITIVE LOGITS

byg

1.02

 your

0.98

 Пусть

0.97

 there

0.96

 others

0.91

be

0.91

not

0.90

 denote

0.87

باشد

0.85

no

0.83

Activations Density 0.025%