INDEX

Explanations

references to extremist groups and hate-related content, especially related to the Ku Klux Klan

oai_token-act-pair · gpt-3.5-turbo

references to hate groups and associated terms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Downloadha

-0.81

ded

-0.77

dra

-0.77

kj

-0.76

*/(

-0.75

neau

-0.75

gob

-0.74

ochond

-0.74

til

-0.73

abilities

-0.72

POSITIVE LOGITS

 Klux

1.23

 Klan

1.12

KKK

0.95

 Sabha

0.81

 affiliation

0.77

 affili

0.74

 NAACP

0.73

 robes

0.73

 Beir

0.72

Jr

0.71

Activations Density 0.011%

references to extremist groups and hate-related content, especially related to the Ku Klux Klan

references to hate groups and associated terms

No Comments

No Known Activations