INDEX

Explanations

negative variations of the verb "to do."

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

 powers

-1.56

à¯ģ

-1.55

 Parenthood

-1.50

à®¿

-1.48

Ĺ

-1.46

itatively

-1.45

ETHOD

-1.44

 increases

-1.41

AN

-1.38

ļ

-1.38

POSITIVE LOGITS

't

2.35

nie

1.84

ned

1.75

ovan

1.75

ÃŃt

1.66

keys

1.50

itÃł

1.48

heit

1.46

jÃ¤

1.42

vette

1.41

Activations Density 0.123%

negative variations of the verb "to do."

No Comments

No Known Activations