INDEX

Explanations

mentions of the word "Crazy" at various intensities

oai_token-act-pair · gpt-3.5-turbo

references to items or concepts associated with "crazy" or "jumbo."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rique

-0.83

rik

-0.78

arers

-0.77

ness

-0.77

ochond

-0.75

lon

-0.75

lÃ©

-0.74

opl

-0.73

rait

-0.70

riz

-0.70

POSITIVE LOGITS

GGGGGGGG

0.75

birds

0.73

SHIP

0.72

emort

0.71

BALL

0.69

igslist

0.68

 slic

0.68

 certific

0.66

gebra

0.64

bird

0.64

Activations Density 0.066%

mentions of the word "Crazy" at various intensities

references to items or concepts associated with "crazy" or "jumbo."

No Comments

No Known Activations