INDEX

Explanations

content related to medical procedures and protocols

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

Ļ

-2.72

¿½

-2.54

ı

-2.45

Ĩ

-2.29

ĳ

-2.28

¨

-2.27

º

-2.20

½

-2.17

ķ

-2.15

¿

-2.14

POSITIVE LOGITS

 exists

1.65

1.61

 within

1.59

 before

1.51

 controls

1.45

 absor

1.42

ahead

1.41

 wins

1.40

ilities

1.38

msgstr

1.37

Activations Density 2.337%

content related to medical procedures and protocols

No Comments

No Known Activations