INDEX

Explanations

This neuron fires on the word “table,” i.e. instances referring to a table or tabular data.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

भाई

0.68

Архі

0.61

 写真

0.61

 fotoğraf

0.56

 पिक्च

0.55

Bis

0.54

 tagline

0.54

keun

0.54

lafen

0.54

Bis

0.54

POSITIVE LOGITS

 table

2.81

表格

2.45

 Table

2.41

table

2.36

 таблицы

2.34

 табли

2.33

Table

2.31

 tables

2.30

テーブル

2.30

 테이블

2.27

Activations Density 0.243%