INDEX

Explanations

contextual transitions

Thank you

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 palettes

0.59

 biscuits

0.58

och

0.57

 pouches

0.56

 canisters

0.54

 הק

0.54

 reeks

0.54

 narrates

0.54

إ

0.53

 הח

0.53

POSITIVE LOGITS

if

0.57

ETT

0.55

 Clemson

0.54

できる

0.52

UU

0.52

ED

0.52

EM

0.51

EF

0.51

我们需要

0.51

 Barton

0.51

Activations Density 0.602%