INDEX

Explanations

income tax

The neuron activates on standalone numeric values (especially floating‐point 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

 继续访问

0.80

</em>

0.58

uintes

0.57

</strong>

0.55

Loksatta

0.55

 огне

0.55

 फायर

0.54

जोरी

0.52

 intellect

0.51

tu

0.50

POSITIVE LOGITS

<b>

0.74

<h1>

0.72

0.68



0.63



0.59



0.57



0.55

<s>

0.54



0.54

<sub>

0.53

Activations Density 0.042%