INDEX

Explanations

references to the name "Ig" or words associated with "Ig"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.0.hook_mlp_out

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

?>

-1.76

thood

-1.65

pite

-1.60

illance

-1.53

 neut

-1.52

 beating

-1.50

etry

-1.49

FIG

-1.41

aying

-1.38

"?"

-1.37

POSITIVE LOGITS

ILITY

1.60

ģ

1.45

ĻĤ

1.44

gens

1.40

lical

1.39

^**

1.39

)|$(

1.36

(@

1.34

alore

1.32

 contexts

1.32

Activations Density 0.007%

references to the name "Ig" or words associated with "Ig"

No Comments

No Known Activations