INDEX

Explanations

likert scale

The neuron activates on any occurrence of the character sequence “lik” (e.g. in “Likert,” “likely,” “kalikrein,” 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

on

-3.13

-2.83

↵

-2.77

龉

-2.34

鬟

-2.34

 Other

-2.31

 Globally

-2.30

 Whether

-2.23

 Somewhat

-2.23

 Sometime

-2.22

POSITIVE LOGITS

lik

3.09

‌های

2.72

࿆

2.69

 они

2.61

？

2.36

τητα

2.34

ίες

2.33

一の

2.33

2.31

neſs

2.25

Activations Density 0.001%