INDEX

Explanations

start with

The neuron strongly activates on the verb “start” (and its conjugated forms like “started,” “starting,” or “starts”).

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

<bos>

-1.77

それが

-1.33

栳

-1.28

we

-1.27

-1.19

这也是

-1.18

-1.16

但却

-1.14

それも

-1.13

POSITIVE LOGITS

 with

1.67

 from

1.40

 afresh

1.33

Ótimo

1.25

,”

1.25

ADUATE

1.21

下雨

1.20

 anew

1.17

 Geraadpleegd

1.16

 Schülern

1.16

Activations Density 0.049%