INDEX

Explanations

terms related to individuals named "Pr" with a high activation value of 9 or 10

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "Pr" followed by numbers or related terms, indicating references to specific individuals or categories

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

 EntityItem

-0.75

INESS

-0.72

ãĥ¼ãĥ«

-0.66

beard

-0.66

 spirited

-0.65

chnology

-0.65

dayName

-0.65

manship

-0.64

 localization

-0.63

rities

-0.63

POSITIVE LOGITS

atche

1.23

udence

1.15

imes

1.06

ima

1.00

inter

0.97

ussia

0.96

icer

0.95

ices

0.95

ager

0.94

ag

0.93

Activations Density 0.010%

terms related to individuals named "Pr" with a high activation value of 9 or 10

instances of the abbreviation "Pr" followed by numbers or related terms, indicating references to specific individuals or categories

No Comments

No Known Activations