INDEX

Explanations

file paths and locations

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 retira

-0.77

Neden

-0.75

xhr

-0.75

 Metode

-0.74

 bouteille

-0.73



-0.73

quakes

-0.72

garh

-0.71

Fakta

-0.70

Metode

-0.70

POSITIVE LOGITS

 path

5.03

path

4.03

 paths

3.97

路径

3.86

 location

3.58

Path

3.50

 Path

3.41

 pathway

3.22

PATH

3.03

 Paths

2.98

Activations Density 0.255%