INDEX

Explanations

references to discriminatory actions or behavior based on various factors such as religion, gender, or sexual orientation

oai_token-act-pair · gpt-3.5-turbo

terms related to discrimination, particularly in social and religious 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

money

-0.83

creation

-0.80

Reb

-0.80

examination

-0.73

Money

-0.70

uro

-0.68

uild

-0.68

anol

-0.67

oly

-0.66

forum

-0.66

POSITIVE LOGITS

 discriminated

1.20

 discriminate

1.15

 discrim

1.08

 discriminating

0.95

inately

0.94

inates

0.92

 retali

0.81

dinand

0.80

 retaliate

0.79

 discriminatory

0.76

Activations Density 0.011%

references to discriminatory actions or behavior based on various factors such as religion, gender, or sexual orientation

terms related to discrimination, particularly in social and religious contexts

No Comments

No Known Activations