INDEX

Explanations

technical methods and techniques

This neuron activates on mentions of specialized scientific or experimental methods, models, and measurement techniques.

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

 jakiś

-1.22

出門

-1.22

さっき

-1.18

ことがある

-1.17

つづく

-1.14

凄く

-1.13

事です

-1.13

のではないか

-1.13

マジで

-1.12

のではない

-1.10

POSITIVE LOGITS

 method

2.20

 approach

2.05

 technique

1.84

<bos>

1.73

 system

1.54

 methods

1.48

禰

1.34

 móds

1.33

 procedure

1.27

วิธี

1.26

Activations Density 0.273%