INDEX

Explanations

specific numerical references, likely in legal or formal contexts

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/2-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.2.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

(%

-1.54

 everywhere

-1.49

-1.47

 alike

-1.44

 hover

-1.42

 experiment

-1.40

 dominate

-1.40

ipel

-1.39

 trillion

-1.38

 pill

-1.37

POSITIVE LOGITS

ETHOD

1.79

unnumbered

1.56

 Strickland

1.55

isciplinary

1.53

IAL

1.50

 COPYRIGHT

1.49

MENTS

1.47

sidered

1.45

VERTISEMENT

1.44

Brad

1.42

Activations Density 0.230%

specific numerical references, likely in legal or formal contexts

No Comments

No Known Activations