INDEX

Explanations

summary table

The neuron is primarily detecting numeric tokens (e.g. standalone digits, years, page numbers, and reference marker numbers).

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

 আর্কাইভ

0.59

>`;

0.59

 واي

0.59

ümüz

0.59

ظم

0.58

 सिंपली

0.57

öm

0.57

ဝ်

0.57

”(

0.56

երը

0.55

POSITIVE LOGITS

<table>

1.36

<h2>

1.00

<h3>

0.73

Normdaten

0.65

|}

0.63

"],"

0.63

<a>

0.62

Przeczytaj

0.61

 frowned

0.61

फ़ी

0.59

Activations Density 0.016%