INDEX

Explanations

formal identifiers and metadata related to programming or documentation

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

 numbers

-1.47

:`

-1.47

Verlag

-1.44

istor

-1.32

inger

-1.30

ische

-1.29

liness

-1.27

ken

-1.26

 number

-1.25

ipzig

-1.25

POSITIVE LOGITS

together

1.46

 deserve

1.43

reements

1.42

 warranted

1.34

while

1.33

 suppression

1.31

Ļª

1.31

 cascade

1.30

 together

1.30

 lane

1.29

Activations Density 5.837%

formal identifiers and metadata related to programming or documentation

No Comments

No Known Activations