INDEX

Explanations

proper nouns, including names of individuals and companies

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

iphate

-0.83

 "$:/

-0.81

Downloadha

-0.79

wcsstore

-0.71

à¼

-0.70

amia

-0.68

tif

-0.64

illas

-0.64

geries

-0.63

agne

-0.63

POSITIVE LOGITS

 founder

1.51

 Director

1.50

 director

1.49

CEO

1.48

 Chairman

1.47

 chairman

1.46

 Chief

1.42

 chief

1.39

 deputy

1.36

 Deputy

1.34

Activations Density 0.436%

proper nouns, including names of individuals and companies

No Comments

No Known Activations