INDEX

Explanations

numeric, byte, pdf, $

The neuron strongly activates on tokens that begin new sentences (i.e. sentence-initial words).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

0.87

й

0.82

hit

0.79

ي

0.78

get

0.76

help

0.76

ি

0.75

0.74

let

0.74

ante

0.73

POSITIVE LOGITS

 oficialmente

0.89

 desempenho

0.88

 fasteners

0.85

 airfoil

0.84

 decisiones

0.81

 carcasses

0.80

 silage

0.79

 dwellings

0.78

 своё

0.78

 dirigentes

0.78

Activations Density 0.000%