INDEX

Explanations

experiment

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

 experiment

-1.39

 Experiment

-1.38

experiment

-1.28

Experiment

-1.20

 EXPERIMENT

-1.12

 experimentation

-1.09

EXPERIMENT

-1.05

 experimental

-1.03

 Experiments

-1.02

 experiments

-1.00

POSITIVE LOGITS

 morire

0.49

 credere

0.43

 poichè

0.41

 papà

0.40

出來的

0.38

 privadas

0.38

 żel

0.38

 sarebbero

0.37

 raiſ

0.37

ība

0.37

Activations Density 0.005%