INDEX

Explanations

This neuron fires on personal narrative or directive utterances—i.e. on first‐ and second‐person pronouns (“I,” “you,” “someone”) together with their main verbs.

phrases or terms

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 SUBMIT

-1.04

 instruk

-1.02

مین

-0.96

Ľ

-0.94

寫

-0.93

 ISLAM

-0.86

 bereit

-0.84

même

-0.82

 экспери

-0.81

avar

-0.80

POSITIVE LOGITS

 می‌گو

1.01

 Wellen

0.93

 Maske

0.88

buhan

0.85

ᱼ

0.84

 فقط

0.83

 сумка

0.82

or

0.82

 Werken

0.82

vedať

0.82

Activations Density 0.058%