INDEX

Explanations

contractions of "is," "are," and "have."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ģ

-2.54

Ģ

-2.44

ł

-2.41

ĸ

-2.25

¡

-2.23

ľ

-2.21

Ł

-2.17

£

-2.08

ĭ

-1.95

Ŀ

-1.95

POSITIVE LOGITS

 gotta

2.25

 been

1.92

 okay

1.84

 gonna

1.82

 gotten

1.77

OK

1.59

 worth

1.57

 bleeding

1.53

 alright

1.45

got

1.43

Activations Density 0.355%

contractions of "is," "are," and "have."

No Comments

No Known Activations