INDEX

Explanations

instances where words are repeated in different contexts within the same document

oai_token-act-pair · gpt-3.5-turbo

the conjunction "and" along with related connections indicating shared experiences or relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

rio

-0.79

oward

-0.79

uto

-0.73

UAL

-0.68

ventory

-0.67

uts

-0.66

UG

-0.66

etting

-0.65

astical

-0.65

ivating

-0.65

POSITIVE LOGITS

 therefore

1.41

 hence

1.27

 thus

1.18

 consequently

1.17

 moreover

1.04

 secondly

1.03

 furthermore

1.00

 cannot

0.86

 possesses

0.83

rogens

0.83

Activations Density 0.312%

instances where words are repeated in different contexts within the same document

the conjunction "and" along with related connections indicating shared experiences or relationships

No Comments

No Known Activations