INDEX

Explanations

second person address

The neuron activates on second-person pronouns (forms of “you”) across different languages.

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

)$

-2.92

</h4>

-2.69

↵

-2.47

</strong>

-2.42

</sup>

-2.28

ኧ

-2.14

ła

-2.08

）

-2.06

 небольшой

-1.88

</em>

-1.87

POSITIVE LOGITS

᯾

2.72

 Tonnen

2.16

—.

2.03

 eigentliche

2.03

 genaue

2.03

ühnen

1.95

ти

1.95

嘥

1.94



1.92

The

1.91

Activations Density 0.003%