INDEX

Explanations

terms related to medical conditions and injuries

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.1.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

dk

-1.63

 same

-1.57

fully

-1.54

 greatest

-1.48

 favor

-1.48

 toast

-1.42

kered

-1.39

htra

-1.39

 adopt

-1.38

 marry

-1.38

POSITIVE LOGITS

antine

1.76

iston

1.69

us

1.63

isms

1.59

thouse

1.55

Stage

1.53

stan

1.52

azol

1.50

Ãªt

1.50

ism

1.50

Activations Density 0.062%

terms related to medical conditions and injuries

No Known Activations