INDEX

Explanations

references to racial identity, specifically focusing on mentions of the term "whites."

oai_token-act-pair · gpt-3.5-turbo

references to the term "whites" in 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Brig

-0.75

bid

-0.71

yrinth

-0.68

ESCO

-0.67

ulous

-0.65

Allow

-0.65

ANK

-0.64

QB

-0.64

 Stra

-0.64

rolog

-0.63

POSITIVE LOGITS

 whites

1.26

 supremacists

0.99

lucent

0.89

paces

0.88

ervative

0.87

 supremacist

0.86

pace

0.86

\\\\\\\\

0.84

peed

0.83

 suprem

0.82

Activations Density 0.006%

references to racial identity, specifically focusing on mentions of the term "whites."

references to the term "whites" in various contexts

No Comments

No Known Activations