INDEX

Explanations

multiple user-related interactions and agreements

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

¿

-1.69

º

-1.57

¼

-1.53

gers

-1.49

ı

-1.44

 crew

-1.43

¯

-1.38

ollow

-1.36

 greatly

-1.33

gradle

-1.33

POSITIVE LOGITS

 recess

1.60

 ahead

1.39

 projections

1.33

quo

1.25

 experience

1.24

ulture

1.24

experience

1.23

 Pence

1.20

 identities

1.20

ation

1.18

Activations Density 2.388%

multiple user-related interactions and agreements

No Comments

No Known Activations