INDEX

Explanations

problem and solution

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

 onData

-0.62

matchCondition

-0.59

kta

-0.53

to

-0.47

ximadamente

-0.46

huy

-0.45

viewBox

-0.45

 cause

-0.44

 invokingState

-0.43

POSITIVE LOGITS

 насељу

0.72

 faſt

0.66

 Jefus

0.64

رشف

0.62

 pleaſure

0.61

 ſever

0.61

 poffible

0.60

 Anſ

0.60

 chofe

0.60

Activations Density 0.005%