INDEX

Explanations

phrases indicating intent or desire, particularly involving "to" followed by verbs

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

nikov

-1.69

anese

-1.61

steps

-1.58

 steps

-1.54

 Statutes

-1.51

teenth

-1.50

 Steps

-1.49

ories

-1.49

âĢĲ

-1.49

 coats

-1.49

POSITIVE LOGITS

 treat

1.77

 resume

1.74

 restore

1.67

 receive

1.65

 guarantee

1.64

 recreate

1.56

 safely

1.54

ren

1.54

 capture

1.52

 remove

1.52

Activations Density 0.090%

phrases indicating intent or desire, particularly involving "to" followed by verbs

No Comments

No Known Activations