INDEX

Explanations

The neuron detects occurrences of words derived from the root “decor,” i.e. terms related to decoration or decorating.

decorating and decoration

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-2.95

⎚

-2.53

It

-2.23

 abject

-2.08

 необходимо

-2.03

derdag

-2.00

end

-1.98

oacute

-1.95

tafogo

-1.94

leq

-1.89

POSITIVE LOGITS

2.33

‘

2.19

芣

2.14

 fashioned

2.00

 filósofo

1.95

 sencillas

1.95

 tradicionais

1.95

1.91

疍

1.88

 mój

1.80

Activations Density 0.008%