INDEX

Explanations

Let's look at

The neuron fires on the conversational transition “Let’s …” constructions—especially “Let’s take a look (at…)” and similar “Let’s [verb] …” invitation phrases.

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

 unangemessene

-1.39

ſelves

-1.39

 largement

-1.36

 dovre

-1.36

 will

-1.35

PENUTUP

-1.30

 bellissimo

-1.29

ਲ਼

-1.28

ך

-1.28

 soggiorno

-1.27

POSITIVE LOGITS

뻗

1.43

<td>

1.34

 some

1.31

zości

1.27

 wzglę

1.27

 bijvoorbeeld

1.27

銲

1.26

.$,

1.24

 just

1.22

jarse

1.21

Activations Density 0.040%