INDEX

Explanations

following conversational turns

This neuron fires on personal pronouns (especially “I” and “you”), i.e. direct first‐ and second‐person references.

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

1.08

0.96

 variously

0.94

0.93

("

0.86

0.85

';

0.82

 needn

0.82

—

0.82

 chiefly

0.81

POSITIVE LOGITS

 piedras

1.15

 اخر

1.13

 dañ

1.08

 नही

1.05

 hacen

1.05

 bisogna

1.05

 hace

1.05

썻

1.05

 haciendo

1.03

 salario

1.01

Activations Density 0.218%