INDEX

Explanations

continuous improvement

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

haltet

-0.94

 aktiviert

-0.90

artiges

-0.86

Pik

-0.86

tissimo

-0.86

inous

-0.85

bno

-0.84

={({

-0.83

🪙

-0.82

zzata

-0.82

POSITIVE LOGITS

 improvement

6.88

 improvements

6.34

 improve

6.25

 improved

5.78

 improving

5.66

 Improvement

5.38

improvement

5.22

improve

5.16

 improves

5.13

 Improvements

5.06

Activations Density 0.294%