INDEX

Explanations

words related to racial issues

oai_token-act-pair · gpt-3.5-turbo

references to racial bias and disparities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

uden

-0.91

icular

-0.87

tower

-0.86

ertodd

-0.85

erva

-0.80

dra

-0.79

hower

-0.78

kens

-0.76

ATURE

-0.76

arent

-0.75

POSITIVE LOGITS

 slurs

1.14

ized

0.99

 minorities

0.95

 violence

0.91

ethnic

0.91

 profiling

0.89

 affili

0.88

 purity

0.87

 animosity

0.86

 backgrounds

0.85

Activations Density 0.015%

words related to racial issues

references to racial bias and disparities

No Comments

No Known Activations