INDEX

Explanations

mentions of the color white

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "white."

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

Closure

-0.74

FUN

-0.71

 Veter

-0.70

Riy

-0.68

orrow

-0.68

lov

-0.68

omics

-0.67

oti

-0.67

conom

-0.65

Aki

-0.64

POSITIVE LOGITS

 white

3.42

white

2.83

 whites

2.68

White

2.32

 WHITE

2.29

 black

2.28

 White

2.03

 yellow

1.86

 whit

1.84

 Whites

1.84

Activations Density 0.026%

mentions of the color white

occurrences of the word "white."

No Comments

No Known Activations