INDEX

Explanations

explanation, reference, or information requests

The neuron fires on prominent topic words that occur in headings or titles (e.g. “verbal,” “CAPTCHA,” “information,” “Results,” “fertilizers,” “video,” “map,” etc.).

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

iblichen

-1.15

nā

-1.12

⤵

-1.08

∓

-1.07

、「

-1.07

{|

-1.03

левая

-1.02

、“

-1.02

蒂

-1.01

mathbb

-1.00

POSITIVE LOGITS

of

1.63

要么

1.29

 already

1.21

ửa

1.21

変わらず

1.19

 vielen

1.16

 either

1.13

勿論

1.13

 agreed

1.13

ごちそう

1.12

Activations Density 0.309%