INDEX

Explanations

mentions of the word "Oscar."

oai_token-act-pair · gpt-3.5-turbo

references to the Oscars and their associated nominations and awards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

bas

-0.71

KY

-0.69

dev

-0.68

 dissolved

-0.68

cit

-0.67

Connect

-0.64

Gaming

-0.64

chron

-0.64

ven

-0.63

hern

-0.62

POSITIVE LOGITS

 Oscar

3.71

 Oscars

2.22

 Emmy

2.20

 Grammy

1.63

 Hugo

1.53

 Nobel

1.38

 Pulitzer

1.37

scar

1.30

 Pixar

1.26

 award

1.21

Activations Density 0.016%

mentions of the word "Oscar."

references to the Oscars and their associated nominations and awards

No Comments

No Known Activations