INDEX

Explanations

the word "spike", particularly in a scientific context

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

 CreateTagHelper

-0.82

RSC

-0.77

 Landis

-0.73

dsl

-0.70

-0.67

 Hous

-0.66

KommentareTeilen

-0.66

RCC

-0.66

Fris

-0.66

 Cress

-0.66

POSITIVE LOGITS

 Spike

1.61

Spike

1.49

spike

1.48

 spike

1.39

 spikes

1.36

spikes

1.31

 spiked

1.21

SPI

1.10

SPI

0.90

iked

0.88

Activations Density 0.005%