INDEX

Explanations

decreasing

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 decreasing

-1.27

reducing

-1.26

 Reducing

-1.23

 reducing

-1.22

Reducing

-1.22

 lessening

-1.13

 lowering

-1.12

 diminishing

-1.10

 Increasing

-1.05

decreasing

-1.04

POSITIVE LOGITS

the

0.66

 speaking

0.53

 their

0.52

gas

0.51

 total

0.50

his

0.49

ang

0.49

kl

0.49

cap

0.48

hil

0.47

Activations Density 0.081%