INDEX

Explanations

know if, why, where, about, how

The neuron strongly activates on the word “know,” especially when it appears in question-like phrases expressing a desire or request for information.

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

-3.11

 mencionado

-1.94

-1.87

 falschen

-1.83

𔘓

-1.81

 одним

-1.80

The

-1.80

 denominado

-1.78

 выборе

-1.75

POSITIVE LOGITS

鮃

2.05

玧

1.92

῞

1.91

︕

1.87

 vlees

1.86

 about

1.81

骤

1.80

猛地

1.78

一闪

1.73

ﮕ

1.70

Activations Density 0.012%