INDEX

Explanations

proper nouns or names

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Willis

-0.94

Mont

-0.88

-0.85

 Mont

-0.79

 Wynne

-0.78

 Truman

-0.76

 Painter

-0.75

 neutron

-0.75

 Polk

-0.73

Mel

-0.72

POSITIVE LOGITS

ig

1.44

igs

1.42

IG

1.40

ags

1.32

ag

1.31

og

1.31

Sag

1.21

Sig

1.18

OG

1.16

AG

1.14

Activations Density 0.352%

proper nouns or names

No Comments

No Known Activations