INDEX

Explanations

execution time order plans

The neuron fires whenever it sees the standalone token “Execution” (or its lowercase form “execution”), i.e. it spots occurrences of that exact word.

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

-3.03

 unſ

-2.38

市の

-2.36

’

-2.34

 あなた

-2.33

鹇

-2.25

-2.22

de

-2.19

 somebody

-2.13

 как

-2.11

POSITIVE LOGITS

<bos>

2.66

ﻂ

2.48

㌧

2.36

责任编辑

2.33

欵

2.30

 These

2.27

噌

2.25

yayım

2.20

owanie

2.16

 gewisser

2.11

Activations Density 0.005%