INDEX

Explanations

mentions of specific names, particularly "Debbie" and "Donna"

oai_token-act-pair · gpt-3.5-turbo

references to individuals, particularly those named Debbie and Donna

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

obar

-0.76

srf

-0.73

¥µ

-0.70

*/(

-0.70

ĻĤ

-0.68

icular

-0.66

Magikarp

-0.64

ibal

-0.64

umn

-0.63

 exploration

-0.63

POSITIVE LOGITS

 Wasserman

1.30

 Debbie

1.01

 DeVos

0.97

 Reynolds

0.91

 Kelley

0.79

 Chal

0.79

 Stab

0.78

 Doll

0.76

 Sachs

0.75

 Ding

0.75

Activations Density 0.009%

mentions of specific names, particularly "Debbie" and "Donna"

references to individuals, particularly those named Debbie and Donna

No Comments

No Known Activations