INDEX

Explanations

The neuron fires on numeric tokens (including integers and 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

No Comments

Negative Logits

1.00

_;

0.68

}}}

0.65

">⁡</

0.61

</sup>

0.59



0.58

<unused2169>

0.58

<unused2194>

0.58

<unused2165>

0.56

POSITIVE LOGITS

<strong>

4.28

<b>

4.03

mathbf

2.41

**

2.02

 mathbf

1.81

<u>

1.56

boldsymbol

1.42

**

1.41

**.

1.40

Activations Density 1.470%