examples and explanations

The neuron fires on extremely common, high‐frequency function words (e.g. “it’s,” “the,” “we,” “should,” etc.).

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

<unused2172>

1.13

<unused2223>

1.08

<unused2130>

1.03

<unused32>

1.02

<unused2123>

1.02

<unused3>

0.98

𝗷

0.96

𝐜

0.96

𝓲

0.96

<unused458>

0.95

POSITIVE LOGITS

1.48

1.00

0.86

0.83

0.82

0.81

0.80

Activations Density 0.005%