INDEX

Explanations

references to gender or gender-related terms, particularly focusing on the term "male"

oai_token-act-pair · gpt-3.5-turbo

mentions of male and female categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

•Gender Features

No Comments

Negative Logits

rieg

-0.79

bley

-0.79

leans

-0.78

heet

-0.77

lay

-0.75

ovie

-0.75

hemy

-0.74

ingen

-0.73

weet

-0.73

Deal

-0.72

POSITIVE LOGITS

volent

1.75

 genital

1.28

vol

0.99

 anatomy

0.91

 genitals

0.89

 infertility

0.87

 counterparts

0.87

 circumcision

0.86

 reproductive

0.85

 gaze

0.85

Activations Density 0.068%

references to gender or gender-related terms, particularly focusing on the term "male"

mentions of male and female categories

No Comments

No Known Activations