INDEX

Explanations

terms related to racism

oai_token-act-pair · gpt-3.5-turbo

references to racism and racially charged comments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

fulness

-0.75

 agility

-0.69

ren

-0.67

riage

-0.66

Mech

-0.66

 stead

-0.66

endor

-0.65

 Imaging

-0.65

 Olympus

-0.65

availability

-0.65

POSITIVE LOGITS

 racist

3.44

racist

2.65

 racists

2.61

 racially

2.29

 homophobic

2.21

 sexist

2.15

 racism

2.06

 discriminatory

1.98

 racial

1.90

acist

1.81

Activations Density 0.040%

terms related to racism

references to racism and racially charged comments

No Comments

No Known Activations