INDEX

Explanations

references to girls and women

oai_token-act-pair · gpt-3.5-turbo

mentions of "Girls" within various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 convincing

-0.77

utherford

-0.75

SPONSORED

-0.73

 loud

-0.71

 Blumenthal

-0.71

ype

-0.69

 Kaine

-0.67

 shaking

-0.64

 ital

-0.64

ypes

-0.63

POSITIVE LOGITS

 Girls

1.45

Girls

1.35

 Actress

0.91

 Girl

0.90

 Boys

0.90

 Haram

0.87

Apps

0.87

poons

0.86

glers

0.85

 Fighters

0.85

Activations Density 0.017%

references to girls and women

mentions of "Girls" within various contexts

No Comments

No Known Activations